DESC:CRASH AND FALL DETECTION
CROSS REFERENCE TO RELATED APPLICATIONS
The present application claims priority from Indian Provisional Patent Application No. 202321075059 filed on 03/11/2023, the entirety of which is incorporated herein by a reference.
TECHNICAL FIELD
Generally, the present disclosure relates to the field of accident detection and alerting systems. Particularly, the present disclosure relates to a system and method for accident detection and alerting systems for a vehicle.
BACKGROUND
With the rise of autonomous vehicles and traffic intensity, reliable crash detection systems are essential for the safe operation and navigation of the vehicles. Advanced crash detection systems provide dependable crash detection mechanisms and increases consumer confidence in new automotive technologies.
Conventionally, the crash detection systems employ machine learning algorithms with autoencoders and isolation forests. The autoencoders are a type of neural network designed for unsupervised learning and consist of an encoder and, a decoder. The encoder compresses input data into a lower-dimensional representation, and the decoder reconstructs the original input from the compressed representation. During model training, the model learns to minimise the difference between the original input and the reconstructed output, effectively learning the underlying patterns and structure of the normal behaviour. The trained autoencoder is used to reconstruct new data. For a normal driving scenario, the reconstruction error (difference between original input and the reconstructed output) is small. For unusual patterns (such as a crash or a fall), the reconstruction error is large. Consequently, a threshold value based on the reconstruction error provides the severity of the crash.
However, there are certain underlining problems associated with the above-mentioned existing mechanism for crash detection mechanism. For instance, the autoencoders also learn the noise in the training data, leading to overfitting. Consequently, the training model with autoencoder performs accurately on training data but inaccurately on real-time data. Further, the overfitting affects the reconstruction sensitivity, leading to errors in the training model, therefore, affecting the overall crash detection mechanism.
Therefore, there exists a need for a mechanism for detecting crash and/or fall of a vehicle that is efficient and overcomes one or more problems as mentioned above.
SUMMARY
An object of the present disclosure is to provide a system for detecting crash and/or fall event of a vehicle.
Another object of the present disclosure is to provide a method of detecting crash and/or fall of a vehicle.
Yet another object of the present disclosure is to provide a system and method for detecting crash and/or fall event of a vehicle capable of an accurate detection of occurrence of crash and/or fall in a vehicle.
In accordance with a first aspect of the present disclosure, there is provided a system for detecting crash and/or fall event of a vehicle, the system comprises:
- a plurality of sensors;
- a data processing unit coupled with the plurality of sensors, and configured to receive input data from the plurality of sensors; and
- a memory module coupled with data processing unit, and configured to store historical data,
wherein the data processing unit is configured to label the received input data based on the stored historical data.
The system and method for detecting crash and/or fall event of a vehicle, as described in the present disclosure, is advantageous in terms of providing a system with enhanced safety and efficiency for detecting crash and/or fall event of a vehicle. Advantageously, model training via machine learning algorithm enables the vehicle to make instantaneous adjustments in vehicle parameters such as, (but not limited to) speed, steering, and braking in an event of possibility of crash and/or fall, thereby enhancing the overall safety and performance of the vehicle.
In accordance with another aspect of the present disclosure, there is provided a method of detecting crash and/or fall of a vehicle, the method comprising:
- scaling the received input data in a consistent range, via the data processing unit;
- labelling the received input data based on historical data, via data processing unit;
- splitting the labelled input data into a training set and a test set via the data processing unit;
- initiating model training of a training set based on a machine learning algorithm, via the data processing unit; and
- classifying a severity of the crash and/or fall, via the trained model.
Additional aspects, advantages, features, and objects of the present disclosure would be made apparent from the drawings and the detailed description of the illustrative embodiments constructed in conjunction with the appended claims that follow.
It will be appreciated that features of the present disclosure are susceptible to being combined in various combinations without departing from the scope of the present disclosure as defined by the appended claims.
BRIEF DESCRIPTION OF DRAWINGS
The summary above, as well as the following detailed description of illustrative embodiments, is better understood when read in conjunction with the appended drawings. For the purpose of illustrating the present disclosure, exemplary constructions of the disclosure are shown in the drawings. However, the present disclosure is not limited to specific methods and instrumentalities disclosed herein. Moreover, those in the art will understand that the drawings are not to scale. Wherever possible, like elements have been indicated by identical numbers.
Embodiments of the present disclosure will now be described, by way of example only, with reference to the following diagrams wherein:
Figures 1 illustrates a block diagram of a system for detecting crash and/or fall event of a vehicle, in accordance with an embodiment of the present disclosure.
Figure 2 illustrates a flow chart of a method of detecting crash and/or fall of a vehicle, in accordance with another embodiment of the present disclosure.
In the accompanying drawings, an underlined number is employed to represent an item over which the underlined number is positioned or an item to which the underlined number is adjacent. A non-underlined number relates to an item identified by a line linking the non-underlined number to the item. When a number is non-underlined and accompanied by an associated arrow, the non-underlined number is used to identify a general item at which the arrow is pointing.
DETAILED DESCRIPTION
The following detailed description illustrates embodiments of the present disclosure and ways in which they can be implemented. Although some modes of carrying out the present disclosure have been disclosed, those skilled in the art would recognize that other embodiments for carrying out or practicing the present disclosure are also possible.
As used herein, the terms “crash”, “collision”, and “accident” are used interchangeably and refer to a vehicle collision with another object, vehicle, or surface. The crash may occur due to (but not limited to) loss of vehicle control, mechanical failure, adverse conditions, or human error. The severity of the crash primarily depends on speed, weight, and angle of collision of the vehicle. The crash of the vehicle causes a sudden deceleration in the vehicle, leading to damage and potential injuries.
As used herein, the term “sensors” refers to devices that detect and measure various physical parameters of a vehicle, and thereby providing critical data to the vehicle control systems. The sensors play a vital role in ensuring the efficient operation, safety, and performance of the vehicle by monitoring associated surrounding conditions, system states, and operating conditions. Various sensors may include (but not limited to) current sensors, voltage sensors, accelerometers, and wheel speed sensors. Additionally, sensors may also include GPS Sensors, pressure sensors and radar sensors.
As used herein, the terms “data processing unit”, “DPU” and “processing unit” are used interchangeably and refer to a specialized electronic component designed to manage, process, and analyse data. Specifically, in a vehicle, the DPU receives the data from various sensors and processes the data in real-time. Further, based on the processing the DPU makes real-time decisions regarding vehicle operation to enhance the rider experience.
As used herein, the terms “memory module”, and “memory unit” are used interchangeably and refer to a hardware component that stores data temporarily or permanently in a computer or an electronic device. The memory module consists of a circuit board with memory chips to store the data. The memory modules are categorized based on the types of storage. For instance, Random Access Memory (RAM) is a temporary storage that provides fast access to the data. Additionally, various types of memory modules may include (but not limited to) Read-only memory (ROM), flash Memory and cache memory.
As used herein, the term “historical data” refers to past records and information related to vehicle incidents, including details of (but not limited to) crashes, driving conditions, vehicle characteristics, and driver behaviours. The historical data is utilized in training machine learning models to identify patterns, predict future crashes, and enhance safety measures.
As used herein, the term “accelerometer” refers to a device that measures the acceleration forces acting on an object. The accelerometer detects changes in velocity and orientation by sensing the rate of acceleration in one or more directions. Further, the accelerometer detects both dynamic acceleration (caused by movement) and static acceleration (due to gravity). Various types of accelerometers may include (but not limited to) single-axis accelerometers, double- axis accelerometers and multi-axis accelerometers.
As used herein, the term “gyroscope” refers to a device that measures the orientation and angular velocity based on the principles of angular momentum. The gyroscope consists of a spinning rotor mounted on its axis of rotation that maintains a constant reference direction. As any object moves or rotates, the gyroscope detects changes in the object orientation and thereby provides object position in three-dimensional space.
As used herein, the term “magnetometer” to a device that measures the strength and direction of a magnetic field. The magnetometer detects the intensity of magnetic forces and is used to identify the presence of magnetic materials or changes in the magnetic conditions. Further, the magnetometer senses the magnetic field around it and provides to determine the orientation or location of an object in relation to magnetic north or south. Various types of magnetometers may include (but not limited to) fluxgate magnetometers, hall effect magnetometers and Optical magnetometers.
As used herein, the term “speed sensor” refers to a device that measures the speed of a vehicle by detecting the rotation of a wheel or the movement of the vehicle. The speed sensor converts the movement into an electrical signal that is processed by the vehicle control unit. Various types of speed sensors may include (but not limited to) wheel Speed Sensors, transmission speed sensors and GPS Speed Sensors.
As used herein, the terms “scale”, and “scaling” are used interchangeably and refer to the process of transforming the features of a dataset to remain within a specific range or distribution. The scaling of the dataset via standardized or normalized ensures that machine learning algorithms perform faster and thereby contribute efficiently to model training. Further, the scaling allows all the features of the dataset to contribute equally to the training model. Various techniques of the scaling may include (but not limited to) normalization scaling, minimum-maximum scaling and standardization scaling.
As used herein, the term “machine learning algorithm” refers to a set of mathematical models and statistical techniques that enable computing processers to learn and make predictions or decisions based on data. The machine learning algorithms utilize the data to identify patterns, classify information, perform predictions, and adapt to real-time data. Specifically, machine learning algorithms play an important role in detecting and analysing a vehicle crash. Various types of machine learning algorithms may include (but not limited to) regression models, decision trees, random forests, and neural networks.
As used herein, the term “model training” refers to a process that enables a model to make predictions or decisions based on data. The model training involves utilizing a training dataset to adjust the model's parameters to minimize errors in its predictions. The model training includes collecting and organizing data into a training set and providing corresponding label to the data. Further, an appropriate machine learning algorithm is selected based on the nature of the data to train the model. Various types of model training include (but not limited to) supervised Learning, and unsupervised Learning.
In accordance with a first aspect of the present disclosure, there is provided a system for detecting crash and/or fall event of a vehicle, the system comprises:
- a plurality of sensors;
- a data processing unit coupled with the plurality of sensors, and configured to receive input data from the plurality of sensors; and
- a memory module coupled with data processing unit, and configured to store historical data,
wherein the data processing unit is configured to label the received input data based on the stored historical data.
Referring to figure 1, in accordance with an embodiment, there is described a system 100 for detecting crash and/or fall event of a vehicle. The system 100 comprises a plurality of sensors 102 (102A-102N), a data processing unit 104 coupled with the plurality of sensors 102 (102A-102N), and a memory module 106 coupled with data processing unit 104, and configured to store historical data. Further, the data processing unit 104 is configured to label the received input data based on the stored historical data.
The labelling of the received input data based on the historical data ensures that the labels are accurately aligned with the previous event data. Consequently, the labelling leads to more accurate and real-time data for model training and analysis. Further, the historical data provides temporal patterns that enhance the model’s understanding of current data and improve the predictive capabilities of the training model. Furthermore, the data processing unit 104 splits the labelled input data into a training data set and a test data set. Advantageously, the splitting of the labelled input data provides a separate set for test data and thereby reduces the overfitting of the training data set. Furthermore, the data processing unit 104 trains the model by utilizing the training data set, via a machine learning algorithm. Consequently, the training model adjusts its parameters based on the training data set, providing an improved model performance as the model learns from patterns and relationships within the training data set. Furthermore, the data processing unit 104 compares a real-time input data with the consecutive event data, for controlling at least one vehicle parameter. The comparison of the real-time data with historical events allows for instantaneous adjustments in vehicle parameters such as, (but not limited to) speed, steering, and braking, thereby enhancing the overall safety and performance of the vehicle.
In an embodiment, the plurality of sensors 102 (102A-102N) comprises at least one accelerometer, at least one gyroscope, at least one magnetometer, and at least one speed sensor. Beneficially, the accelerometer and gyroscope provide changes in acceleration and angular velocity respectively, and thereby enable advanced stability control systems to maintain vehicle balance and traction. Further, the magnetometer provides vehicle accurate positioning and direction to enable real-time adjustments in vehicle systems to enhance stability during cornering, acceleration, or sudden manoeuvres.
In an embodiment, the data processing unit 104 is configured to scale the received input data in a consistent range. The scaling of the received input data enables all the features of the input data to contribute equally during training, allowing machine learning algorithms to converge faster and perform efficiently. Consequently, the training models are trained quickly as no adjustment for varying scales of the input data is required. Further, scaling of the received input data reduces numerical instability during computations and thereby provides a more reliable and stable training process.
In an embodiment, the data processing unit 104 is configured to split the labelled input data into a training data set and a test data set, wherein the training data set and the test data set comprise consecutive events data. Advantageously, the splitting of the labelled input data provides a separate set for test data and thereby, reduces the overfitting of the training data set. Further, the separation of the labelled input data enables the evaluation of the training model on unseen data that provides an accurate simulation of the model performance. Furthermore, the consecutive events data provides the temporal sequence and context of data that is utilized for time-dependent analysis of the training model.
In an embodiment, the data processing unit 104 is configured to initiate model training by utilizing the training data set, via a machine learning algorithm. Advantageously, the consistent training of the model based on the training data set ensures that the training models are trained uniformly and thereby enhances the training model prediction efficiency. Further, the training model adjusts its parameters based on the training data set, providing an improved model performance as the model learns from patterns and relationships within the training data set.
In an embodiment, the data processing unit 104 is configured to validate the trained model by utilizing test data set. The validation of the training model enables the data processing unit 104 to efficiently identify the overfitting condition of the training model. Further, the data processing unit 104 is configured to make real-time adjustments in the training model based on the overfitting condition. Furthermore, based on the validation result, the data processing unit 104 performs validation of the input data allowing for correction or improvement in the training data collection process.
In an embodiment, the data processing unit 104 is configured to classify a severity of the crash and/or fall, via the trained model. The trained model is configured to assess the severity of a crash or fall based on incoming sensor data, allowing for immediate response actions. Further, the trained model provides a clear categorization of the severity of the crash such as, minor, moderate, severe, and thereby, aiding in risk assessment and response planning. Furthermore, classifying the crash severity provides real-time data for analysing trends and patterns in the crash, facilitating improvements in vehicle design and safety features.
In an embodiment, the data processing unit 104 is configured to compare a real-time input data with the consecutive event data, for controlling at least one vehicle parameter. The comparison of the real-time data with historical events allows for instantaneous adjustments in vehicle parameters such as, (but not limited to) speed, steering, and braking, thereby enhancing overall safety and performance. Further, the data collected through real-time comparisons is utilized in refining the machine learning models, and thereby improving the model accuracy and responsiveness in future iterations.
In accordance with a second aspect, there is described method 200 of detecting crash and/or fall of a vehicle, the method 200 comprises:
- scaling the received input data in a consistent range, via the data processing unit 104;
- labelling the received input data based on historical data, via data processing unit 104;
- splitting the labelled input data into a training set and a test set via the data processing unit 104;
- initiating model training of a training set based on a machine learning algorithm, via the data processing unit 104;and
- classifying a severity of the crash and/or fall, via the trained model.
Figure 2 describes a method of detecting crash and/or fall of a vehicle. The method 200 starts at a step 202. At the step 202, the method comprises scaling the received input data in a consistent range, via the data processing unit 104. At a step 204, the method comprises labelling the received input data based on historical data, via data processing unit 104. At a step 206, the method comprises splitting the labelled input data into a training set and a test set via the data processing unit 104. At a step 208, the method comprises initiating model training of a training set based on a machine learning algorithm, via the data processing unit 104. At a step 210, the method comprises classifying a severity of the crash and/or fall, via the trained model. The method 200 ends at the step 210.
In an embodiment, the method 200 comprises scaling the received input data in a consistent range, via the data processing unit 104.
In an embodiment, the method 200 comprises splitting the labelled input data into a training data set and a test data set, wherein the training data set and the test data set comprises consecutive events data, via the data processing unit 104.
In an embodiment, the method 200 comprises initiating model training by utilizing the training data set, via a machine learning algorithm, via the data processing unit 104.
In an embodiment, the method 200 comprises validating the trained model by utilizing test data set, via the data processing unit 104.
In an embodiment, the method 200 comprises classifying a severity of the crash and/or fall, by the trained model, via the data processing unit 104.
In an embodiment, the method 200 comprises comparing a real-time input data with the consecutive event data, for controlling at least one vehicle parameter, via the data processing unit 104.
In an embodiment, the method 200 comprises scaling the received input data in a consistent range, via the data processing unit 104. Furthermore, the method 200 comprises splitting the labelled input data into a training data set and a test data set, wherein the training data set and the test data set comprises consecutive events data, via the data processing unit 104. Furthermore, the method 200 comprises initiating model training by utilizing the training data set, via a machine learning algorithm, via the data processing unit 104. Furthermore, the method 200 comprises validating the trained model by utilizing test data set, via the data processing unit 104. Furthermore, the method 200 comprises classifying a severity of the crash and/or fall, by the trained model, via the data processing unit 104. Furthermore, the method 200 comprises comparing a real-time input data with the consecutive event data, for controlling at least one vehicle parameter, via the data processing unit 104.
In an embodiment, the method 200 comprises scaling the received input data in a consistent range, via the data processing unit 104. Furthermore, the method 200 comprises labelling the received input data based on historical data, via data processing unit 104. Furthermore, the method 200 comprises splitting the labelled input data into a training set and a test set via the data processing unit 104. Furthermore, the method 200 comprises initiating model training of a training set based on a machine learning algorithm, via the data processing unit 104. Furthermore, the method 200 comprises classifying a severity of the crash and/or fall, via the trained model.
Based on the above-mentioned embodiments, the present disclosure provides significant advantages such as (but not limited to) enhanced efficiency for detecting crash and/or fall event of a vehicle, and real-time adjustments in vehicle parameters based on model training, thereby, ensuring overall safety and performance of the vehicle.
It would be appreciated that all the explanations and embodiments of the system 100 also apply mutatis-mutandis to the method 200.
In the description of the present invention, it is also to be noted that, unless otherwise explicitly specified or limited, the terms “disposed,” “mounted,” and “connected” are to be construed broadly, and may for example be fixedly connected, detachably connected, or integrally connected, either mechanically or electrically. They may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Modifications to embodiments and combinations of different embodiments of the present disclosure described in the foregoing are possible without departing from the scope of the present disclosure as defined by the accompanying claims. Expressions such as “including”, “comprising”, “incorporating”, “have”, and “is” used to describe and claim the present disclosure are intended to be construed in a non-exclusive manner, namely allowing for items, components or elements not explicitly described also to be present. Reference to the singular is also to be construed to relate to the plural where appropriate.
Although embodiments have been described with reference to a number of illustrative embodiments thereof, it should be understood that numerous other modifications and embodiments can be devised by those skilled in the art that will fall within the spirit and scope of the principles of this disclosure. More particularly, various variations and modifications are possible in the component parts and/or arrangements of the subject combination arrangement within the scope of the present disclosure, the drawings, and the appended claims. In addition to variations and modifications in the component parts and/or arrangements, alternative uses will also be apparent to those skilled in the art.
,CLAIMS:WE CLAIM:
1. A system (100) for detecting crash and/or fall event of a vehicle, the system (100) comprises:
- a plurality of sensors (102);
- a data processing unit (104) coupled with the plurality of sensors (102), and configured to receive input data from the plurality of sensors (102); and
- a memory module (106) coupled with data processing unit (104), and configured to store historical data,
wherein the data processing unit (104) is configured to label the received input data based on the stored historical data.

2. The system (100) as claimed in claim 1, wherein the plurality of sensors (102) comprises at least one accelerometer, at least one gyroscope, at least one magnetometer, and at least one speed sensor.

3. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to scale the received input data in a consistent range.

4. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to split the labelled input data into a training data set and a test data set, wherein the training data set and the test data set comprises consecutive events data.

5. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to initiate model training by utilizing the training data set, via a machine learning algorithm.

6. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to validate the trained model by utilizing test data set.

7. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to classify a severity of the crash and/or fall, via the trained model.

8. The system (100) as claimed in claim 1, wherein the data processing unit (104) is configured to compare a real-time input data with the consecutive event data, for controlling at least one vehicle parameter.

9. A method (200) of detecting crash and/or fall of a vehicle, the method (200) comprises:
- scaling the received input data in a consistent range, via the data processing unit (104);
- labelling the received input data based on historical data, via data processing unit (104);
- splitting the labelled input data into a training set and a test set via the data processing unit (104);
- initiating model training of a training set based on a machine learning algorithm, via the data processing unit (104); and
- classifying a severity of the crash and/or fall, via the trained model.

10. The method (200) as claimed in claim 9, the method (200) comprises validating the test set based on the model training of the training set via the data processing unit (104).

11. The method (200) as claimed in claim 9, the method (200) comprises comparing a real-time input data with the consecutive event data, for controlling at least one vehicle parameter via the data processing unit (104).