Claims:We Claim:
1. A method of contactless braking detection for a vehicle, comprising:
obtaining a plurality of accelerometer and gyroscope samples from an inertial measurement unit (103) at periodical time intervals;
dynamically calculating resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s (103) mounting angle;
calculating, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of each of the plurality of accelerometer samples in the forward direction is less than zero, and (ii) a window average that calculates a running average of incoming accelerometer samples in the forward direction;
storing the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
calculating the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein a moving window size parameter indicates the number of accelerometer samples by which a next window moves forward from a current window;
comparing the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
turning OFF a brake indicator (108), through a network (106) if (a) the values of the negative count are lower than the predetermined thresholds or (b) the values of the window average are higher than the predetermined thresholds; else
comparing the values of the plurality of gyroscope samples that are obtained for the set of windows with corresponding predetermined thresholds to check any vehicle turn false scenario, and when the vehicle turn false scenario is observed, then weighing down an intensity of the brake signal; and
turning ON the brake indicator (108), through a network (106), with the intensity that is proportional to the values of the window average, which are indicative of the intensity of braking.

2. The method as claimed in claim 1, comprising
coupling the inertial measurement unit (103) and a microcontroller (102) with a frame of the vehicle at a fixed mounting angle.

3. The method as claimed in claim 1, wherein the sampling rate, the window size, the moving window size, the set of windows, and the predefined thresholds for the negative count and the window average for braking detection are determined and adapted based on at least one of (i) a vehicle type, (ii) a construction variant within the vehicle type, (iii) a placement of the inertial measurement unit (103), (iv) a terrain, or (v) an end user application requirement for timing and accuracy.


4. The method as claimed in claim1, wherein the window size is less than 500 milliseconds, the negative count and the window average are calculated for at least 1 window, and the moving window size is at least 1 sample.

5. The method as claimed in claim 4, wherein the predefined threshold for the negative count is at least 50 percent of the samples in the window, and the predefined threshold for each of the window average is less than or equal to -0.15 m/s2.

6. The method as claimed in claim 5, comprising varying the intensity of the brake indicator (108) based on (i) low level deceleration, (ii) medium level deceleration and (iii) high level deceleration of the vehicle.

7. The method as claimed in claim 6, comprising
calibrating an accelerometer sensor (104) to determine the mounting angle of the accelerometer sensor (104), comprising the steps of:
having the vehicle on a flat surface and in a non-moving state;
sampling the accelerometer sensor (104) for a predefined time period; and
obtaining the accelerometer samples to calculate a resultant roll and a resultant pitch, wherein the resultant
roll=tan^(-1)??(a_y/a_z )? and pitch=tan^(-1)??(a_x/v(?a_y?^2+ ?a_z?^2 ))?
wherein ax (702) represents accelerometer samples in a roll axis, ay (704) represents accelerometer samples in a pitch axis, and az (706) represents accelerometer samples in a yaw axis.


8. The method as claimed in claim 7, comprising
calculating a resultant yaw using at least one of (i) measurement through a system such as a protractor and entering the value into the system, (ii) using a gyroscope, wherein the gyroscope is initially hold in a zero yaw in a heading direction and then turning the gyroscope until the gyroscope is aligned to the actual mounting yaw, and (iii) using a magnetometer, wherein calculating magnetometer values from two positions (a) heading angle and (b) mounting angle and calculating their difference.


9. The method as claimed in claim 8, comprising calculating a dynamic acceleration of the vehicle in the forward direction using the equation:
a_f=a_y cos??? cos???- a_z ? ? sin??? cos???-a_x ? ? sin??? sin???- a_x ? ? cos??? cos??? sin?? ? ?
wherein ?=roll,?=pitch,and ?=yaw.

10. A system of contactless braking detection for a vehicle using the method as claimed in claim 1, comprising a microcontroller (102), an inertial measurement unit (103), and a brake indicator (108) connected to a network (106).

11. A system of contactless braking detection for a vehicle comprising
a sensing device (101) that is coupled to a frame of the vehicle at a fixed mounting angle, wherein the sensing device (101) comprises an Inertial Measurement Unit (IMU) (103), and a microcontroller (102) that is communicatively coupled to the inertial measurement unit (103), wherein the inertial measurement unit (103) comprises a 3-axis accelerometer sensor (104) and a 3-axis gyroscope, wherein the microcontroller (102) comprises
an IMU samples obtaining unit (202) that obtains a plurality of 3-axis accelerometer and 3-axis gyroscope samples from an inertial measurement unit (103) at periodical time intervals;
a forward axis acceleration calculation unit (206) that dynamically calculates resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s (103) mounting angle;
a dynamic feature calculation unit (208) that calculates, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of the incoming sample is less than zero, and (ii) a window average, which is a running average of incoming samples;
a feature values storing unit (210) that stores the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
a moving window unit (212) that calculates the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein the moving window size parameter indicates the number of accelerometer samples by which the next window moves forward from the current window;
a features comparison unit (214) that compares the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
a vehicle turn false removal unit (216) that compares the values of the gyroscope that are obtained for the set of windows with corresponding predetermined thresholds to check any false positive scenario occurring due to vehicle turning, and when such a false scenario is observed, then weighing down an intensity of the brake signal; and
a brake indication communication unit (218) that sends an indication to a brake indicator (108), through a network (106), wherein the brake indicator (108) is turned ON when (a) the values of the negative count are higher than their predetermined thresholds and (b) the values of the window average are lower than the predetermined thresholds, with the intensity that is proportional to the values of the window average which are indicative of the intensity of braking, else the brake indicator (108) is turned OFF.


12. The system as claimed in claim 11, comprising an accelerometer sensor calibration (204) unit that calibrates the accelerometer sensor (104) to determine the mounting angle of the inertial measurement unit (103).


13. The system as claimed in claim 12, wherein the calibration of the inertial measurement unit (103) using the accelerometer sensor calibration unit (204) comprises the steps of:
having the vehicle on a flat surface and in a non-moving state;
sampling the accelerometer sensor (104) for a predefined time period; and
obtaining the accelerometer samples to calculate a resultant roll and a resultant pitch, wherein the resultant
roll=tan^(-1)??(a_y/a_z )? and pitch=tan^(-1)??(a_x/v(?a_y?^2+ ?a_z?^2 ))?
wherein ax (702) represents accelerometer samples in a roll axis, ay (704) represents accelerometer samples in a pitch axis, and az (706) represents accelerometer samples in a yaw axis.


14. The system as claimed in claim 12, wherein the forward axis acceleration calculation unit (206) calculates the dynamic acceleration of the vehicle in the forward direction using an equation of:
a_f=a_y cos??? cos???- a_z ? ? sin??? cos???-a_x ? ? sin??? sin???- a_x ? ? cos??? cos??? sin?? ? ?
wherein ?=roll,?=pitch,and ?=yaw.
, Description:BACKGROUND
Technical Field
The embodiments herein generally relate to braking detection and, more particularly, to a system and method of contactless braking detection for a vehicle.
Description of the Related Art
As the number of vehicles on roads increases, especially during rush hours, the potential for accidents also increases. A lot of accidents are attributed to drivers not being able to adequately assess how fast or how suddenly vehicles in front of them are braking. Braking warning devices have been described in the art to provide an indication when vehicles in front are braking by a variety of different mechanisms, in order to potentially reduce accidents.
Contact based braking detection is found in most motor vehicles which indicates whenever a brake lever is pressed. However, there are scenarios when a vehicle brakes without the apparent brake lever mechanism, such as when engine braking, downshifting etc. Also, not all kinds of vehicles have a brake lever based braking mechanism. Hence, contact based braking detection does not encompass the entire spectrum of braking scenarios across different kinds of vehicles.
Contactless braking approaches have been developed to overcome the limitations of contact braking detection. Earlier work on contactless braking detection was based on mechanical sensors such as inertia switches, pendulums etc. to detect decelerations of vehicles. However, those sensors were quite bulky, expensive and not easy to customize across vehicle types.
Recent advancements in electronics and sensing technologies have enabled smaller and cheaper form factors of sensing devices. These have resulted in newer areas of applications such as Internet of Things (IoT) which focuses on low-power, smart, connected devices.
Some recent approaches on contactless braking detection use Global Positioning System (GPS) receivers to determine a vehicle’s speed. Other approaches use magnetic wheel speed sensors to determine the vehicle’s speed. In either of these approaches, whenever the vehicle speed reduces beyond a certain threshold, a brake light is turned on. However, GPS is a power hungry sensing technique and is also expensive. Hence, it is not suitable as yet, for a ubiquitous speed sensor, especially for IoT applications. Magnetic wheel speed sensors require user intervention in setting up on each vehicle and cannot be moved from one vehicle to another with relative ease.
Micro Electromechanical Systems (MEMS) based Inertial Measurement Unit (IMU) sensors are suitable to be mounted across various vehicles to detect their decelerations due to their smaller form factor, low power consumption and low cost. They can also be interfaced with relative ease with a microcontroller. The MEMS sensor based brake detection devices can be clamped to a vehicle’s frame to detect its decelerations and can be detached and attached to different vehicles very quickly.
Most approaches based on MEMS sensors based braking detection for motor vehicles first estimate the sensor mounting position, then filter the incoming accelerometer data to remove vibrational noise and compare the forward axis acceleration values against preset thresholds.
There has been little work done specifically on bicycle braking which is prone to more vibrations and noise than other motor vehicles. For a sensor that can be mounted to any rigid part of the vehicle frame including a bicycle handlebar, which undergoes more movement and vibrations than other parts of the vehicle, a filter and threshold method wouldn’t suffice to detect brakes of all intensities within a good timing accuracy.
Accordingly, there remains a need for an improved approach for detecting braking accurately with improved timing.
SUMMARY
The present disclosure provides a method of contactless braking detection for a vehicle, the method including steps of
obtaining a plurality of accelerometer and gyroscope samples from an inertial measurement unit at periodical time intervals;
dynamically calculating resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s mounting angle;
calculating, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of each of the plurality of accelerometer samples in the forward direction is less than zero, and (ii) a window average, which is a running average of incoming accelerometer samples in the forward direction;
storing the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
calculating the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein the moving window size parameter indicates the number of accelerometer samples by which the next window moves forward from the current window;
comparing the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
turning OFF a brake indicator, through a network if (a) the values of the negative count are lower than the predetermined thresholds or (b) the values of the window average are higher than the predetermined thresholds; else
comparing the values of the plurality of gyroscope samples that are obtained for the set of windows with corresponding predetermined thresholds to check any vehicle turn false scenario, and when the vehicle turn false scenario is observed, then weighing down an intensity of the brake signal; and
turning ON the brake indicator, through a network, with the intensity that is proportional to the values of the window average, which are indicative of the intensity of braking.
The present disclosure also provides a system of contactless braking detection for a vehicle, comprising
a sensing device that is coupled to a frame of the vehicle at a rigid mounting angle, wherein the sensing device comprises an Inertial Measurement Unit (IMU), and a microcontroller that is communicatively coupled to the inertial measurement unit, wherein the inertial measurement unit comprises of a 3-axis accelerometer sensor and a 3-axis gyroscope, wherein the microcontroller comprises
an IMU samples obtaining unit that obtains a plurality of 3-axis accelerometer and 3-axis gyroscope samples from an inertial measurement unit at periodical time intervals;
a forward axis acceleration calculation unit that dynamically calculates resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s mounting angle;
a dynamic feature calculating unit that calculates, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of the incoming sample is less than zero, and (ii) a window average, which is a running average of incoming samples;
a feature values storing unit that stores the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
a moving window unit that calculates the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein the moving window size parameter indicates the number of accelerometer samples by which the next window moves forward from the current window;
a features comparison unit that compares the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
a vehicle turn false removal unit that compares the values of the gyroscope that are obtained for the set of windows with corresponding predetermined thresholds to check any false positive scenario occurring due to vehicle turning, and when such a false scenario is observed, then weighing down an intensity of the brake signal; and
a brake indication communication unit that sends an indication to a brake indicator, through a network, wherein the brake indicator is turned ON when (a) the values of the negative count are higher than their predetermined thresholds and (b) the values of the window average are lower than the predetermined thresholds, with the intensity that is proportional to the values of the window average which are indicative of the intensity of braking, else the brake indicator is turned OFF.
These and other aspects of the embodiments herein will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions, while indicating preferred embodiments and numerous specific details thereof, are given by way of illustration and not of limitation. Many changes and modifications may be made within the scope of the embodiments herein without departing from the spirit thereof, and the embodiments herein include all such modifications.

BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
FIG. 1 illustrates a system view of a contactless braking detection system according to an embodiment herein;
FIG. 2 illustrates an exploded view of a microcontroller of FIG. 1 according to an embodiment herein;
FIGS. 3A-3F are graphical representations illustrating acceleration in a riding direction, values of the features negative count (Feature 1) and average value of acceleration per window (Feature 2), ground truth for braking obtained from a pressure sensor mounted on the brake lever and braking detection calculated using the method as described in the summary and in detail in the following sections according to an embodiment herein;
FIGS. 4A-4B illustrate a process of contactless braking detection of FIG. 1 for the vehicle according to an embodiment herein;
FIG. 5 illustrates an exemplary view of a sensing device of FIG. 1 that is mounted on various locations of a bicycle frame according to an embodiment herein;
FIGS. 6A-6C depict the sequence of events that occur when a brake is applied in a bicycle according to an embodiment herein;
FIGS. 7A-7C illustrate tabular views of sensor data obtained from an accelerometer sensor of FIG. 1, a hall effect sensor for measuring vehicle speed and the pressure sensor according to an embodiment herein; and
FIGS. 8A-8B illustrate a method of contactless braking detection for the vehicle according to an embodiment herein.

DETAILED DESCRIPTION OF THE INVENTION
The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
The present disclosure provides a method of contactless braking detection for a vehicle, the method comprising steps of
obtaining a plurality of accelerometer and gyroscope samples from an inertial measurement unit at periodical time intervals;
dynamically calculating resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s mounting angle;
calculating, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of each of the plurality of accelerometer samples in the forward direction is less than zero, and (ii) a window average, which is a running average of incoming accelerometer samples in the forward direction;
storing the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
calculating the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein a moving window size parameter indicates the number of accelerometer samples by which a next window moves forward from a current window;
comparing the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
turning OFF a brake indicator, through a network if (a) the values of the negative count are lower than the predetermined thresholds or (b) the values of the window average are higher than the predetermined thresholds; else
comparing the values of the plurality of gyroscope samples that are obtained for the set of windows with corresponding predetermined thresholds to check any vehicle turn false scenario, and when the vehicle turn false scenario is observed, then weighing down an intensity of the brake signal; and
turning ON the brake indicator, through a network, with the intensity that is proportional to the values of the window average, which are indicative of the intensity of braking.
The method improves the timing and accuracy of contactless braking detection. The method effectively detects deceleration of the vehicle to turn the brake indicator on or off and avoids false detections. The method has been devised and iterated through rigorous data collected for different types of braking scenarios to improve the timing and accuracy of braking detection. The method detects braking independent of the mounting angle of the inertial measurement unit due to the calibration based mounting angle calculation. Although majority of the data has been described in relation to bicycles, the method is suitable for application on other motor vehicles such as electric scooters, motor bikes, cars and other vehicles which have similar deceleration gradients on braking.
The method focuses on detecting braking in vehicles with high accuracy and sub-second timing of detection in a wide range of terrains and vehicle types. The method proposes a feature based detection algorithm that is not very computationally expensive which is deployed on to low power sensors. In one embodiment, the method is suitable for being used on wearable IoT devices which have limited on board energy and computational power. The method detects brakes of all intensities with sub-second timing accuracy. The method calculates two features from riding axis acceleration data for multiple overlapping data windows. The method eliminates vehicle turn false scenarios through gyroscope values. The method eventually compares the values of the two features against their preset thresholds to detect braking.
In an embodiment, the method may be tuned for better accuracy or timing of braking detection as there is a trade-off between timing and accuracy. In a preferred embodiment, the method provides a balance between accuracy and timing of detection.
In one embodiment, the method classifies average decelerations less than or equal to -0.25 meter per second square (m/s2) occurring within 480 milliseconds time window as braking. The method may also calculate features (e.g. a negative count and a window average) for a window of samples (e.g. 30 samples) by employing the moving window technique. The method may further consider feature values from multiple consecutive windows to turn the brake indicator on or off.
In one embodiment, the outputs of the brake indicators may be latched to stay on for a few milliseconds after a braking condition has been met. This is a configurable and optional feature.
In one embodiment, brake pressing timing is calculated through a pressure sensor mounted on the brake levers of a vehicle. In some vehicles (e.g. bicycle, motorcycle etc.), there is a gap between the brake pads and the wheels, so the vehicle does not start decelerating as soon as the brake levers are pressed, but only after the brake pads start rubbing against the rim. Since the accelerometer sensor senses the actual deceleration of the vehicle, the method response begins only after the actual deceleration (i.e. speed reduction) starts resulting in a slight delay between pressing the levers and the brake detection.
For training the algorithm or the method, a wide variety of data associated with braking of the vehicle has been collected. The data includes but it is not limited to, (a) a type of vehicle (e.g. bike, motorbike, car), (b) a construction variant within the vehicle type (e.g. a mountain bike, a road bike or a hybrid bike), (c) a type of terrain (e.g. a tar road, a cement road or a mud road), (d) a type of incline (e.g. uphill, downhill and flat roads), (e) braking intensities (e.g. hard braking, medium braking, low braking, pulsing brake and speed change brake), (f) mounting angles of the inertial measurement unit (e.g. from -90° to 90° Roll angle ) and (g) road disturbances (e.g. potholes, bumps, uneven roads etc.).
In one embodiment, the method determines dynamic changes in riding terrains including uphill and downhill inclines, which result in changing the resultant riding axis. Road inclination may be dynamically measured by combining other sensing modalities (e.g. altimeter sensor, GPS etc.) to correct for the inclination.
Even without any additional sensor fusion from inclination sensors, the method can be applied to highly steep road gradients (e.g., 35%) and the detection accuracy remains unaffected with inclination changes. In a few cases, timing might be altered from the corresponding flat state ride, resulting in a slight delay on a positive gradient (e.g. uphill) and a slight advancement in detection on negative gradient (e.g. downhill).
According to an embodiment, the method further includes coupling the inertial measurement unit and a microcontroller with a frame of a vehicle at a fixed mounting angle.
The inertial measurement unit and the microcontroller may be coupled to the vehicle at various locations such as a handle bar, a saddle post etc. as long as the inertial measurement unit has a fixed mounting angle. However, certain locations of the inertial measurement unit may result in better sensor data as compared to others. For example, the inertial measurement unit located close to the center of gravity of the vehicle experiences lesser vibrations than when located on the handlebar which experiences more movement and vibrations.
According to an embodiment, wherein the sampling rate, the window size, the moving window size, the set of windows, and the predefined thresholds for the negative count and the window average for braking detection are determined and adapted based on at least one of (i) a vehicle type, (ii) a construction variant within the vehicle type, (iii) a placement of the inertial measurement unit, (iv) a terrain, or (v) an end user application requirement for timing and accuracy.
End user applications may have different weightage on timing and accuracy. In some applications, timing may be more critical than accuracy, i.e. false detections do not cause as much damage as delayed detections would. In other cases, accuracy could play a bigger role than detection timing. The algorithm or the method can be customized for achieving the required end application goals for timing and accuracy. The thresholds for the two features along with the number of windows may be adapted for achieving these requirements. Having fewer windows’ features enables faster detections than having more windows for detection which result in slightly delayed but more accurate detections. The feature negative count keeps a check on noise so having a lower threshold for this enables faster detections but possibly more false detections. Having a higher threshold on window average feature results in more accuracy with slightly delayed detection and vice versa.
According to one embodiment, the window size is less than 500 milliseconds, the negative count and the window average are calculated for at least one window, the moving window size is at least 1 sample. The predefined threshold for the negative count is at least 50 percent of the samples in the window, and the predefined threshold for each of the window average is less than or equal to -0.15 m/s2. The sampling rate of the sensing device is at least 50 Hertz (Hz). The predetermined time period for sampling the accelerometer sensor for calibration is at least 1 second.
For a balance between timing and accuracy, the window size may be less than 400 milliseconds, the negative count and the window average may be calculated for 3 windows or 4 windows, and the moving window size may be less than or equal to 25 percent of the window size, , the predefined thresholds for the negative count may be 60-80 percent of the samples in the window, the predefined thresholds for each of the window average may be less than or equal to -0.25 m/s2, and the sampling rate of the sensing device may be at least 100 Hertz (Hz).
In certain rides, additional data from pressure sensor and hall effect sensor is also obtained. The pressure sensor is mounted on the brake pad to provide a reference for braking timing. The hall effect sensor data is used to calculate the vehicle speed.
According to yet another embodiment, the method further includes
calibrating the accelerometer sensor to determine the mounting angle of the accelerometer sensor, comprising the steps of:
having the vehicle on a flat surface and in a non-moving state;
sampling the accelerometer sensor for a predefined time period; and
obtaining the accelerometer samples to calculate a resultant roll and a resultant pitch, wherein the resultant
roll=tan^(-1)??(a_y/a_z )? and pitch=tan^(-1)??(a_x/v(?a_y?^2+ ?a_z?^2 ))?
wherein ax represents accelerometer samples in a roll axis, ay represents accelerometer samples in a pitch axis, and az represents accelerometer samples in a yaw axis.
According to yet another embodiment, the method further includes calculating a resultant yaw using at least one of (i) measurement through a system such as a protractor and entering the value into the system, (ii) using a gyroscope, wherein the gyroscope is initially hold in a zero yaw in a heading direction and then turning the gyroscope until the gyroscope is aligned to the actual mounting yaw, and (iii) using a magnetometer, wherein calculating magnetometer values from two positions (a) heading angle and (b) mounting angle and calculating their difference.
According to an embodiment, the method further includes calculating a dynamic acceleration of the vehicle in the forward direction using the equation:
a_f=a_y cos??? cos???- a_z ? ? sin??? cos???-a_x ? ? sin??? sin???- a_x ? ? cos??? cos??? sin?? ? ?
wherein ?=roll,?=pitch,and ?=yaw.
According to yet another embodiment, the method includes varying an intensity of the brake indicator based on (i) low level deceleration, (ii) medium level deceleration and (iii) high level deceleration of the vehicle.
The method not only enables granularity in the brake detection but also helps in weighing down any false detections. The braking detections may be further classified to obtain higher granularity. In one embodiment, the intensity is proportional to the values of the window average. The intensity indicates an extent of the vehicle deceleration.
Depending on the vehicle type and the inertial measurement unit mounting position, certain false detections may occur. In vehicles (e.g. bicycles), three major types of false detections for sensor mounted on handlebar: i) sharp turns which result in change of heading direction ii) occasional false occurrences when riding the vehicle in a standing position (for e.g., during an uphill incline), which results in rotation around riding axis iii) extremely bad roads and irregular terrains which result in high amplitude vibrational noise along with occasional handlebar movement.
The gyroscope data is used to mitigate false occurrences due to conditions i) sharp turns, ii) occasional false occurrences in cases of handlebar movements occurring in condition, and iii) extremely bad roads and irregular terrains. Handlebar turning results in the gyroscope values changing across the rotational axis. Similarly, during standing and riding a bike, gyroscope values along the riding axis change. The system uses the gyroscope data markers to avoid false occurrences of braking.
In another embodiment, the present disclosure provides a system of contactless braking detection for a vehicle using the method as described above, the system including a microcontroller, an inertial measurement unit and a brake indicator.
The present disclosure further provides a system of contactless braking detection for a vehicle includes
a sensing device that is coupled to a frame of the vehicle at a fixed mounting angle, wherein the sensing device comprises an Inertial Measurement Unit (IMU), and a microcontroller that is communicatively coupled to the inertial measurement unit, wherein the inertial measurement unit comprises a 3-axis accelerometer sensor and a 3-axis gyroscope, wherein the microcontroller comprises
an IMU samples obtaining unit that obtains a plurality of 3-axis accelerometer and 3-axis gyroscope samples from an inertial measurement unit at periodical time intervals;
a forward axis acceleration calculation unit that dynamically calculates resultant acceleration in forward direction for each of the plurality of accelerometer samples based on the inertial measurement unit’s mounting angle;
a dynamic feature calculation unit that calculates, for each value of acceleration in the forward direction (af): (i) a negative count, whose value increments if magnitude of the incoming sample is less than zero, and (ii) a window average, which is a running average of incoming samples;
a feature values storing unit that stores the values of the negative count and the window average when the plurality of accelerometer samples is equal to a size parameter of a window, wherein the window is a cluster of the plurality of accelerometer samples;
a moving window unit that calculates the negative count and the window average for a plurality of windows until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique, wherein the moving window size parameter indicates the number of accelerometer samples by which the next window moves forward from the current window;
a features comparison unit that compares the values of the negative count and the window average that are obtained for the set of windows with corresponding predetermined thresholds;
a vehicle turn false removal unit that compares the values of the gyroscope that are obtained for the set of windows with corresponding predetermined thresholds to check any false positive scenario occurring due to vehicle turning, and when such a false scenario is observed, then weighing down an intensity of the brake signal; and
a brake indication communication unit that sends an indication to a brake indicator, through a network, wherein the brake indicator is turned ON when (a) the values of the negative count are higher than their predetermined thresholds and (b) the values of the window average are lower than the predetermined thresholds, with the intensity that is proportional to the values of the window average which are indicative of the intensity of braking, else the brake indicator is turned OFF.
The advantages of the system are thus identical to those disclosed above in connection with the present method and the embodiments listed above in connection with the method apply mutatis mutandis to the system. The inertial measurement unit may be coupled to the frame of a vehicle to obtain acceleration data of the vehicle periodically. In one embodiment, the inertial measurement unit is coupled to a handle bar of the vehicle. The sensing device may be Micro-Electro-Mechanical Systems (MEMS) Inertial Measurement Unit (IMU). The vehicle may be, but not limited to, a bicycle, a two-wheeler vehicle, a three-wheeler vehicle, a four-wheeler vehicle and the like.
The microcontroller may obtain accelerometer samples (e.g. accelerometer data) from the inertial measurement unit at periodic time intervals.
The microcontroller processes the accelerometer samples to communicate an indication (e.g. a brake signal) to the brake indicator through the network. The network may be, but not limited to: an internet, a wide area network, a wired cable network, a broadcasting network, a wired communication network, a wireless communication network, a fixed wireless network, a mobile wireless network, Bluetooth and the like. The indication may include information to turn the brake indicator on at various intensities or turning it off. In one embodiment, the brake indicator includes a light to indicate other riders coming behind the vehicle when the rider is braking. The brake indicator may include a speaker to alarm a sound to alert other riders coming behind the vehicle when the rider is braking. The brake indicator, if it is light weight, may be placed on a rider backpack or anywhere on the rear side of the vehicle for visibility to riders behind.
According to an embodiment, the system further includes an accelerometer sensor calibration unit that calibrates the inertial measurement unit to determine the mounting angle of the inertial measurement unit.
According to another embodiment, the calibration of the accelerometer sensor using the accelerometer sensor calibration unit includes the steps of:
having the vehicle on a flat surface and in a non-moving state;
sampling the accelerometer sensor for a predefined time period; and
obtaining the accelerometer samples to calculate a resultant roll and a resultant pitch, wherein the resultant
roll=tan^(-1)??(a_y/a_z )? and pitch=tan^(-1)??(a_x/v(?a_y?^2+ ?a_z?^2 ))?
wherein ax represents accelerometer samples in a roll axis, ay represents accelerometer samples in a pitch axis, and az represents accelerometer samples in a yaw axis.
The roll axis refers to a rotation of the sensing device around X-axis of the inertial measurement unit. The pitch refers to a rotation of the sensing device around Y-axis of the inertial measurement unit. The yaw axis refers to a rotation of the sensing device around Z-axis of the inertial measurement unit.
According to another embodiment, wherein the forward axis acceleration calculation unit calculates the dynamic acceleration of the vehicle in the forward direction using the equation:
a_f=a_y cos??? cos???- a_z ? ? sin??? cos???-a_x ? ? sin??? sin???- a_x ? ? cos??? cos??? sin?? ? ?
wherein ?=roll,?=pitch,and ?=yaw.
Embodiments of the present disclosure may eliminate false detections occurring in existing approaches and improve on the timing of contactless braking detection of the vehicle. The embodiments of the present disclosure have been determined through rigorous data collection across various types of braking scenarios and analysis to provide improved accuracy and timing of braking detection of the vehicle.

DETAILED DESCRIPTION OF THE DRAWINGS
Referring now to the drawings and more particularly to FIGS. 1 through 8B, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.
FIG. 1 illustrates a system view of a contactless braking detection system 100 according to an embodiment herein. The contactless braking detection system 100 includes a sensing device 101, a network 106 and a brake indicator 108. The sensing device 101 includes a microcontroller 102 and an inertial measurement unit 103. The inertial measurement unit 103 includes an accelerometer sensor 104 and a gyroscope sensor 105. The functions of these parts have been described above.
FIG. 2 illustrates an exploded view of the microcontroller 102 of FIG. 1 according to an embodiment herein. The microcontroller 102 includes an IMU samples obtaining unit 202, an accelerometer sensor calibration unit 204, a forward axis acceleration calibration unit 206, a dynamic features calculation unit 208, a feature values storing unit 210, a moving window unit 212, a features comparison unit 214, a vehicle turn false removal unit 216, and a brake indication communication unit 218. The functions of these parts have been described above.
FIGS. 3A-3F are graphical representations illustrating acceleration in a riding direction, values of the features, negative count (Feature 1) and average value of acceleration per window (Feature 2), ground truth for braking obtained from a pressure sensor mounted on the brake lever and braking detection calculated using the method as described in the summary and in detail in the following sections according to an embodiment herein.
FIG. 3A: In this scenario, a rider was riding a city bicycle on a tar road at a speed of ~19 kilometer per hour (km/h) and applied medium intensity brake by pressing both the brake levers simultaneously. The brake levers are pressed close to sample number 3087 (corresponding to 30.87 seconds since the start of ride). Typically, the resistance of the pressure sensor reduces when the brake levers are pressed. Hence a dip in the pressure sensor values corresponds to a brake applied. The higher the magnitude of braking, the higher the change in resistance from non-pressed state. The graphical representations of the feature 1 (e.g. the negative count) and the feature 2 (e.g. the window average) depict a clear change while the rider is braking. The feature 1 values increase and the feature 2 values decrease in magnitude. The first instance of detection from the method appears at sample number 3132 corresponding to 31.32 seconds in time since the start of the ride. It is to be noted that the instance of pressing the brake lever does not instantly indicate deceleration of the bike as there is a gap between the brake pads and the rim. From the observed data, there is no instance of false detection during the ride.
FIG. 3B: In this scenario, the rider was riding a road bicycle on a tar road at a speed of ~33 km/h and applied high intensity brake by pressing the single brake lever present in the road bicycle. The brake lever is pressed close to sample number 3132 corresponding to 31.32 seconds in time since the start of ride. The graphical representations of the feature 1 values (e.g. the negative count) and the feature 2 values (e.g. the window average) depict a clear difference while the rider is braking. The first instance of detection from the method appears at sample number 3156 corresponding to 31.56 seconds since the start of ride.
FIG. 3C: In this scenario, the rider was riding a hybrid bicycle on a tar road at a slow speed ~12 km/h and applied very low intensity brake by pressing both the brake levers simultaneously. The graphical representations of the feature 1 values (e.g. the negative count) and the feature 2 values (e.g. the window average) depict a clear difference while the rider is braking. As the braking intensity is low, the feature 2 values of deceleration as observed in the feature plot are also low in magnitude. The brake levers are pressed close to sample number 2740 corresponding to 27.40 seconds since the start of ride and the first instance of detection occurs at sample number 2838 corresponding to 28.38 seconds in time. It is to be noted that the instance of pressing the brake lever does not instantly indicate deceleration of the bike as there is a gap between the brake pads and the rim.
FIG. 3D: In this scenario, the rider was riding an electric scooter on a tar road at a speed of ~60 km/h and applied high intensity brake by pressing both the brake levers simultaneously. The inertial measurement unit 103 is placed on a dashboard of the electric scooter. The graphical representations of the feature 1 values (e.g. the negative count) and the feature 2 values (e.g. the window average) depict a clear change while the rider is braking. The feature 1 values increase and the feature 2 values decrease in magnitude. The brake levers are pressed at sample number 2796 corresponding to 27.96 seconds since the start of ride and the first instance of detection occurs at sample number 2836 corresponding to 28.36 seconds in time.
FIG. 3E: In this scenario, the rider was riding a 350cc motor bike on a tar road at a speed of ~60 km/h and applied medium intensity brake by pressing the front brake lever and the rear brake pedal simultaneously. The inertial measurement unit 103 is placed on a dashboard of the motor bike. The graphical representations of the feature 1 values (e.g. the negative count) and the feature 2 values (e.g. the window average) depict a clear change while the rider is braking. The feature 1 values increase and the feature 2 values decrease in magnitude. The front brake lever and the rear brake pedal are pressed at sample number 4781 corresponding to 47.81 seconds since the start of the ride and the first instance of detection occurs at sample number 4854 corresponding to 48.54 seconds in time.
FIG. 3F: In this scenario, the rider was riding a car on a tar road at a speed of ~60 km/h and applied very low intensity brake by pressing the brake pedal. The inertial measurement unit 103 is placed near the handbrake of the car. The graphical representations of the feature 1 values (e.g. the negative count) and the feature 2 values (e.g. the window average) depict a clear change while the rider is braking. The feature 1 values increase and the feature 2 values decrease in magnitude. The brake pedal was pressed in the beginning when the car was stationary. During the ride, the brake pedal was pressed at sample number 4131 corresponding to 41.31 seconds since the start of ride and the first instance of detection occurs at sample number 4224 corresponding to 42.24 seconds in time.
FIGS. 4A-4B illustrate a process for contactless braking detection of FIG. 1 for the vehicle according to an embodiment herein. At step 402, the microcontroller 102 and the inertial measurement unit 103 are mounted on the frame of the vehicle at a convenient mounting angle. At step 404, the accelerometer sensor 104 is calibrated to determine the mounting angle while keeping the frame of the vehicle in straight riding position. At step 406, the roll and pitch are calculated based on accelerometer data during calibration. The Yaw, if present, may be measured using other sensors such as gyroscope and magnetometer. At step 408, the accelerometer is sampled at periodical time intervals. At step 410, the acceleration in forward axis is calculated from the accelerometer samples based on the mounting angle. The accelerometer samples are obtained from the pitch axis, the roll axis and the yaw axis. At step 412, the values of the two features are calculated. The two features are (i) an average of acceleration values within a window, and (ii) number of samples with values < 0 within the window. At step 414, it is checked if the moving window size is reached. If yes, then the features values are compared across predetermined thresholds, else sampling continues until the moving window size is reached to a predefined number. At step 416, if the feature values do not match the predetermined thresholds then the brake indicator is turned OFF at step 418, else at step 420 it is checked if the gyroscope values are indicative of a vehicle turn false positive scenario. If a false positive scenario is observed, then the intensity of the brake signal as determined by window average values is weighed down at step 422. At step 424, the brake indicator 108 is turned ON with the intensity that is proportional to the values of the window average which are indicative of the intensity of braking, else the brake indicator is turned OFF at step 418.
FIG. 5 illustrates an exemplary view of the sensing device 101 of FIG. 1 that is mounted on various locations of a bicycle frame according to an embodiment herein. The sensing device 101 may be mounted on various locations (e.g. 502, 504, 506, and 508) such as the handlebar 502 or the saddle post 508 of the bicycle.
FIGS. 6A-6C depict the sequence of events that occur when a brake is applied in a bicycle according to an embodiment herein. A rider applies the brakes of the bicycle at step 602. The accelerometer sensor 104 senses deceleration of the bicycle at step 604, generates an indication (e.g. a brake signal) and transmits the indication via the network. The brake indicator 108 (e.g. lights) on the rider's backpack connected to the sensing device 101 via Bluetooth Low Energy is turned ON at step 606 based on the indication (e.g. a brake signal) received from the sensing device 101.
FIG. 7A illustrates tabular view of sensor data obtained from all axes (e.g. 702, 704, 706) of the accelerometer sensor 104 of FIG. 1. FIGS. 7B-7C depict sensor data obtained from all axes (e.g. 702, 704, 706) of the accelerometer sensor 104 during ride along with a hall effect sensor 710 for measuring vehicle speed and a pressure sensor 708 according to an embodiment herein.
FIGS. 8A-8B illustrate a method for contactless braking detection for the vehicle according to an embodiment herein. At step 802, the plurality of accelerometer and gyroscope samples is obtained from the inertial measurement unit 103 at periodical time intervals. At step 804, the resultant acceleration in forward direction is calculated dynamically for each of the plurality of accelerometer samples based on a mounting angle of the inertial measurement unit 103. At step 806, (i) a negative count whose value increments if magnitude of each of the plurality of accelerometer samples in the forward direction is less than zero and (ii) a window average, which is a running average of incoming accelerometer samples in the forward direction, are calculated.
At step 808, the values of the negative count and the window average are stored when the plurality of accelerometer samples is equal to the size parameter of the window. At step 810, the negative count and the window average for the plurality of windows is calculated until the plurality of windows is equal to a predefined number, as indicated by a parameter set of windows, by applying a moving window technique. The moving window size parameter indicates the number of accelerometer samples by which the next window moves forward from the current window. At step 812, the values of the negative count and the window average that are obtained for the set of windows are compared to corresponding predetermined thresholds. If the feature values do not match the predetermined thresholds, then the brake indicator is turned OFF at step 818, else at step 814 the values of gyroscope that are obtained for the set of windows are compared with corresponding their predetermined thresholds to check for any vehicle turn false positive scenario. If a false scenario is observed, then the intensity of the brake signal is weighed down at step 814. At step 816, the brake indicator 108 is turned ON through the network 106 with intensity that is proportional to the intensity of braking.
The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the appended claims.