DESC:FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION

(See Section 10 and Rule 13)

Title of invention:
SYSTEM AND METHOD FOR DETECTION OF ROAD IRREGULARITIES
APPLICANT:
MICELIO MOTORS PRIVATE LIMITED
Having Address:
No.58, 15th cross, 2nd phase JP Nagar, Bengaluru-560078, India

The following specification particularly describes the invention and the manner which is to be performed.

CROSS-REFERENCE TO RELATED APPLICATIONS AND PRIORITY
The present application claims priority from an Indian patent Application No: 202141058247, filed on 14th of Dec 2021, incorporated herein by a reference.

TECHNICAL FIELD
The present disclosure relates to a system and a method for detection of road irregularities, and more particularly to a system and a method for continuous monitoring, developing and detecting the road surface irregularities using data captured from a two-wheeler.

BACKGROUND
The subject matter discussed in the background section should not be assumed to be prior art merely as a result of its mention in the background section. Similarly, a problem mentioned in the background section or associated with the subject matter of the background section should not be assumed to have been previously recognized in the prior art. The subject matter in the background section merely represents different approaches, which in and of themselves may also correspond to implementations of the claimed technology.

Worldwide and especially, India is rigged with road fatalities. Road irregularities being one of the most prominent reasons of them all. Every stretch of roads is either with potholes or bumpers or speed breakers. Early detection of different road irregularities provides a significant possibility of reducing the chances of accidents thereby saving lives and vehicle damages. A few probable solutions to this problem may be by detecting and identifying roads and warning the drivers while commuting.

Few solutions proposed for detecting road surface irregularities rely on a system comprising mobile phones in combination with accelerometer sensors. These particular systems allowed the user to monitor and detect the road surface anomaly and surface hazards in real time similar to traffic congestion information in navigation applications. However, performing analysis locally on the mobile phone created issues with the battery life and the heating of the mobile since the processor would never be in idle mode throughout the drive.

To resolve the issue of over running of the processor and optimally manage the issue of bandwidth and availability of the network another proposed solution relies on splitting the functionality between the mobile device and a remote server. The pre-processing can be done by the mobile and the processing is done at the remote server. However, the major drawback of the solution is the continuous orientation of the mobile to detect the changes in X, Y and Z direction.

Further, the U.S. Patent No. US10235770B2 discloses a system for pothole detection comprises an input interface configured to receive sensor data and a pothole detector configured to determine a pothole based at least in part on the sensor data using a model, wherein the model is used to classify sensor data; and store pothole data associated with the pothole. In one of its embodiments, the sensor data comprises accelerometer data, gyroscope data, GPS data, or any other appropriate sensor data. Further, the road irregularity (like pothole) definition comprises a minimum sensor data threshold and a maximum sensor data threshold (e.g., when the sensor data is measured to record a reading between the minimum sensor data threshold and the maximum sensor data threshold, it is determined that a pothole has been encountered). Further, a vehicle recorder stores data that is determined by a local processor to be a potential pothole. The potential pothole data is transmitted to a server and further analyzed using a model that determines a likelihood of the potential pothole data being from an actual pothole. In the event that the potential pothole data is likely an actual pothole (e.g., likelihood is greater than a threshold), the potential pothole is confirmed by review of a video. In some embodiments, there is no review to confirm the pothole. In various embodiments, the pothole determination is checked periodically, randomly, or at any other appropriate interval. Further, in some embodiments, time-frequency features, such as the short-time Fourier Transform (STFT) or wavelet decomposition coefficients are also used to describe the waveform. These extracted features are used to encode the shape signature of the potholes. The shape signatures extracted from the example events are then used to train a model of the prototypical pothole signature. The learning of the model is done by a machine learning algorithm (e.g., Support Vector Machines (SVM), Neural Networks or other relevant model classification technique). The drawback of the present application is the entire data is analyzed for road irregularities thereby consuming more processing power, further the same analysis is carried at two location i.e. at the vehicle end and at a remote server, thereby making the process repetitive.

Thus, there is an indispensable need for a system and a method for detecting any kind of road irregularities encompassing potholes, bumpers, and speed-breakers. Further, the detection method and the system are also needed to monitor and train a prediction model that alleviates the aforementioned-technical challenges/drawbacks while also resolving the issue of over running of the processors and optimally managing the issue of bandwidth and availability of the network.

SUMMARY
This summary is provided to introduce concepts related to a system and a method for detection of road irregularities with the help of sensors mounted on a two-wheeler and the concepts are further described below in the detailed description. This summary is not intended to identify essential features of the claimed subject matter nor is it intended for use in determining or limiting the scope of the claimed subject matter.

In an implementation of the present disclosure a method to detect road irregularities is disclosed. In accordance with an exemplary embodiment an acquired data may be obtained from an Inertial Measurement Unit (IMU) sensor by using a data acquisition module. Further geo-coordinates may be mapped on the acquired data by the data acquisition module to generate a first data set. In accordance with the exemplary embodiment filter may be applied using the data acquisition module to segregate anomalies detected in the first data set. Further the segregated first data may be transferred to a server. The segregated first data may further be overlapped on a map corresponding to the geo-coordinates mapped by the data acquisition module.

In another implementation of the present disclosure a system for detection of road irregularities is disclosed. The system may comprise an IMU (Inertial Measurement Unit) sensor positioned at a steering rod of a vehicle. Further a data acquisition module may be communicably coupled to the IMU sensor. The data acquisition module may further comprise a global positioning module embedded into the data acquisition module. The system may further comprise a filter module, configured to run a Butterworth filter, embedded into the data acquisition module. Further a machine learning module may be embedded into the data acquisition module.

BRIEF DESCRIPTION OF DRAWINGS
The detailed description is described with reference to the accompanying Figures. In the Figures, the left-most digit(s) of a reference number identifies the Figure in which the reference number first appears. The same numbers are used throughout the drawings to refer like features and components.

Figure 1(a-d) illustrates a positional layout of sensors in a vehicle in accordance with an embodiment of a present subject matter.

Figure 2 illustrates data modelling in accordance with an embodiment of a present subject matter.

Figure 3 illustrates an architectural system in accordance with an embodiment of a present subject matter.

Figure 4a illustrates a flow chart for the irregularity detection method via sensors on the vehicle in accordance with an embodiment of a present subject matter.

DETAILED DESCRIPTION
Reference throughout the specification to “various embodiments,” “some embodiments,” “one embodiment,” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases “in various embodiments,” “in some embodiments,” “in one embodiment,” or “in an embodiment” in places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner in one or more embodiments.
In accordance with an exemplary embodiment of the present disclosure a method to detect road irregularities is disclosed. In accordance with the embodiment, acquired data is obtained from an Inertial Measurement Unit (IMU) sensor by a data acquisition module. The acquired data may be captured for a vibrational frequency between the range of 1Hz to 100Hz. Acquiring the data may further comprise obtaining derived features by capturing acceleration in X and Z direction of cartesian-coordinate system by the IMU sensor and further selecting mean value, peak value and standard deviation value for X and Z axis from the derived features.

In accordance with the embodiment the acquired data may be mapped with geo-coordinates by the data acquisition module to generate a first data set. Further a filter may further be applied to the first data set using the data acquisition module to segregate anomalies detected in the first data set. For e.g. a Butterworth Filter, may be applied to retrieve anomalies between 1 Hz to 15 Hz. The segregated first data may further be transferred to a server. Further the segregated first data may be overlapped on a map corresponding to the geo-coordinates mapped by the data acquisition module.

In accordance with the exemplary embodiment the segregated first data may be analysed and tagged as potholes, bumps or speed brakers based on attributes of each.

In accordance with another exemplary embodiment of the present disclosure a system for detection of road irregularities may be disclosed. The system may comprise of an IMU (Inertial Measurement Unit) sensor positioned at a steering rod of a vehicle. Further a data acquisition module may be communicably coupled to the IMU sensor. The data acquisition module may be mounted in a cavity provided in the rear of the vehicle, and further a global positioning module may be embedded into the data acquisition module.

Further a filter module embedded into the data acquisition module may be configured to run a Butterworth filter. The filter may be configured to retrieve vibrational frequency between 1Hz to 15 Hz. In accordance with the exemplary embodiment the system may further comprise a machine learning module embedded into the data acquisition module. Further the data acquisition module may comprise a communication module configured to communicably connect the data acquisition module with a remote server.

Referring to Figures 1(a - d), illustrates a two-wheeler with an Inertia Measurement Unit and a Data Acquisition module (OTC) in accordance with an exemplary embodiment. The Inertia Measurement Unit 110 in accordance with the present exemplary embodiment may be mounted on a steering of the two-wheeler 100. The Inertia Measurement Unit 110 (IMU) may be mounted on a cross-member positioned between two forks of a suspension system. The mounting of the (IMU) 110 on the cross-member enables capturing of road vibrations with higher accuracy.
Further in accordance with the exemplary embodiment the IMU 110 may be communicably connected to the Data Acquisition module (OTC) 120. The Data Acquisition module (OTC) 120 may be mounted in a cargo portion or cargo hold of the two-wheeler 100. The cargo hold or cargo portion may be provided under the carriage of the vehicle or under the seat. The data acquisition module 120 in accordance with another exemplary embodiment may be mounted on the chassis of the two-wheeler below the seat assembly. The data acquisition system 120 may further comprise an accelerometer, a processor, communication module and a global positioning module embedded on a circuit board or connected to the board.

Now Figure 2, illustrates a plurality of data modelling modules in accordance with an embodiment of a present subject. The data modelling modules 200 may further comprise about five modules of execution namely: a data collection module 210, data filtering module 220, feature extraction/selection module 230, machine learning module 240, and visualisation module 250.

The at least five modules may be embedded on a single processor and positioned within the OTC. In another exemplary embodiment each modules may be embedded in independent processor, and further mounted at various position on the two-wheeler with communicably connected with each other. Further in yet another exemplary embodiment of the distributed system, few of the modules from the at least five modules may be mounted on two-wheeler and few at a remote location.

The data collection module 210 may be configured to collect data from the plurality of sensors placed on the two-wheeler. In accordance with an exemplary embodiment the data from IMU mounted on the steering rod of the vehicle, accelerometer mounted in a cargo section, or the GPS module mounted on the vehicle is captured by the data collection module 210. The accelerometers may be configured to measure accelerations in X, Y and Z axis respectively. Further, GPS signals may be used to capture the latitude and longitude positions.

The data collected and aggregated by the data collection module 210 may be further filtered using the data filtering module 220. The raw data obtained by the collection module 210 is filtered to remove unwanted data such as road frequencies beyond 15Hz using the data filtering module 220. Filtering techniques such as Butterworth filter may be used to filter the data between 1 and 15 Hz.

In accordance with the exemplary embodiment the feature extraction/selection module 230 may be applied to filtered data identify and segregate different feature/data detected in the filtered data. The features play an important role for distinguishing between potholes and bumps. Raw data contains mainly acceleration in X and Z axis respectively. Further, derived features are obtained from the X and Z accelerations. The derived features are mean, peak , X/Z ratio and standard deviation values of X and Z axis respectively.

Further, the data obtained from feature extraction/selection module 230 is analysed using the machine learning module 240. Machine learning methods such as classification algorithms are used to distinguish between potholes and bumps. In addition, potholes, bumps or speed brakers are labelled as 0, and 1 respectively. Labelling is carried out in order to distinguish between potholes, bumps and straight lines. The labelled data is classified using machine learning algorithms such as K-NN, Decision trees and Random Forest respectively. Thus, the algorithm classifies into potholes, bumps and speed breakers. Further, Vibration Dose Value (VDV) is computed which in determines the ride quality. The machine algorithms which are classified as potholes, bumps or speed breaker is further visualised using the visualisation module 250 by overlapping the bumps and potholes on a google map.

Referring to Figure 3, illustrates an architectural system in accordance with an embodiment of a present subject matter. The architectural layout 300 comprises a vehicle 100. A block 300 as illustrated may comprise an IMU sensor 110, an accelerometer 120a and a GPS module 120b. In accordance with the exemplary embodiment the components of block 300 captures various data pertaining to road irregularities in raw data format. The raw data may further be mapped with geo-coordinates. Further data set generated may be filtered in block 320 the acquired data 310 to extract relevant values. With the help of different features which are extracted after the data processing 320, data labelling 330 phase is initiated. Further, labelling is carried out in order to distinguish between potholes, bumps and straight lines. The labelled data is classified using machine learning algorithms such as K-NN, Decision trees and Random Forest respectively. Thus, the algorithm classifies into potholes, bumps and speed breakers. The classified data may further be transposed or overlapped with a map 350 (for the co-ordinates of different road irregularities) for easy visualisation by the user.

Now, referring Figure 4, a flow chart 400 is further illustrated to provide a structured method and a system to detect, acquire, classify, train, and prevent any kind of road irregularities.

At step S410, a data collection 210 process takes place for acquiring data 310 from the IMU sensors 110, accelerometers 120a, and GPS 120b to generate a data set. The acquired data 310 depends on several different requirements and features. The different requirements for data collection 210 also effect the hardware requirements. These requirements along with the hardware requirements provide selection criteria for the types of data collection methods. The acquired data 310 is captured for a vibrational frequency ranging between 1Hz to 100Hz.

At step S420, a data filtering 220 process takes place under the data processing 320 wherein the raw acquired data 310 is filtered to remove the unwanted data, such as road frequencies beyond a frequency of 15Hz. Further, a filter module embedded into the data acquisition module 120 is configured to run a Butterworth filter to filter the raw acquired data 310 between 1Hz and 15 Hz. Once, the raw acquired data 210 is filtered and anomalies are removed, the refined and segregated first data is transferred to a server.

At step S430, the feature extraction/selection 230 process takes place wherein the features extraction helps distinguishing the road irregularities into potholes, speed-breakers, and bumps or any other road defects. Further, the derived features are obtained from the X and Z accelerations which are selected as mean, peak, X/Z ratio, and standard deviation values of X and Z axes wherein data labelling 330 the segregated data is initiated into 0, 1, and 2 for potholes, bumps or speed brakers, respectively. Labelling is carried out in order to distinguish between potholes, bumps and straight lines.

At step S440, the Machine Learning model/process 240 initiates data classification 340 to analyse the labelled data. The Machine learning model 240 uses algorithms such as classification algorithms for data classification 340 and thereby, distinguish the labelled data between different road irregularities like potholes, speed-breakers, bumpers, and any other types of road irregularities. Further, the data classification 340 with the machine learning model 240 (algorithms such as K-NN, Decision trees and Random Forest) tags the road irregularities into potholes, bumpers or speed brakers based on the attributes of each which were selected at step S430.

At step S450, the visualisation 250 process takes place wherein the segregated, classified and labelled data are overlayed on maps 350 (for coordinates of any kind of road irregularities). Further, the Vibration Dose Value (VDV) is computed which in turn determines the ride quality of the vehicle 100.

The embodiments illustrated above, especially related to the method of classifying the acquired data provides a set of training data as well. Further, the offline data that is collected is subjected to pre-processing techniques and the unwanted data is eliminated. Also, the use of Butterworth filter ensures removal of anomalies (frequencies beyond 15Hz). Filtered data is furthered ahead for the data labelling process wherein the labelled data is classified using machine learning algorithms thereby helping the algorithms get trained for the futuristic prediction models.

Various modifications to the embodiment will be readily apparent to those skilled in the art and the generic principles herein may be applied to other embodiments. However, one of ordinary skill in the art will readily recognize that the present disclosure is not intended to be limited to the embodiments illustrated but is to be accorded the widest scope consistent with the principles and features described herein.

The foregoing description shall be interpreted as illustrative and not in any limiting sense. A person of ordinary skill in the art would understand that certain modifications could come within the scope of this disclosure.

The embodiments, examples and alternatives of the preceding paragraphs or the description and drawings, including any of their various aspects or respective individual features, may be taken independently or in any combination. Features described in connection with one embodiment are applicable to all embodiments, unless such features are incompatible.
,CLAIMS:We claim:
1. A method to detect road irregularities, the method comprises:
(S410), a data collection (210) process by acquiring data (310) from an IMU sensor 110, accelerometers (120a), and a GPS (120b) to generate a data set;
(S420), data filtering (220) process using the data acquisition module (120) under data processing (320) to segregate anomalies detected in the first data set and transferring the segregated first data to a server;
(S430), feature extraction/selection (230) process to distinguish between different types of road irregularities and the derived features are selected further as mean, peak, X/Z ratio, and standard deviation values of X and Z axes and initiating the data labelling (330) process of labelling the potholes, bumps or speed brakers as 0, 1, and 2 respectively;
(S440), Machine Learning Model (240) process for data classification (340) to classify the labelled data into potholes, bumps and speed breakers; and
(S450), visualising (250) or overlaying process for segregating the first data on a map (350) corresponding to the geo-coordinates mapped by the data acquisition module.

2. The method as claimed in claim 1, further comprises capturing the acquired data between a vibrational frequency range of 1Hz to 100Hz.

3. The method as claimed in claim 1 and claim 2, wherein applying filter comprises applying a Butterworth Filter, and retrieving anomalies between 1Hz to 15 Hz.

4. The method as claimed in claim 1, wherein acquiring data (310) further comprises obtaining derived features by capturing acceleration in X and Z direction of cartesian-coordinate system by the IMU sensor (110).

5. The method as claimed in claim 4, further comprises selecting mean value, peak value and standard deviation value for X and Z axis from the derived features.

6. The method as claimed in claim 1, further comprises analysing the first data set using machine learning algorithms (240).

7. The method as claimed in claim 1, wherein the analysing further comprises tagging the road anomalies in the first data set into potholes, bumps or speed brakers based on the attributes of each.

8. A system for detection of road irregularities, the system comprising:
an IMU (Inertial Measurement Unit) sensor (110) positioned at a steering rod of a vehicle (100);
a data acquisition module (120) communicably coupled to the IMU sensor (110), wherein the data acquisition module (120) further comprises a global positioning module (120b) embedded into the data acquisition module (120);
a filter module (220), configured to run a Butterworth filter, embedded into the data acquisition module (120); and
a machine learning module (240) embedded into the data acquisition module (120).

9. The system as claimed in claim 8, wherein the data acquisition module (120) is mounted in a cavity provided in the vehicle (100).

10. The system as claimed in claim 8, wherein the filter module (220) is configured to retrieve vibrational frequencies between 1Hz to 15 Hz.

11. The system as claimed in claim 8, wherein the data acquisition module (120) further comprises a communication module configured to communicably connect the data acquisition module (120) with a remote server.

Dated 14th Day of December 2021