Parkour - A pothole detection application

As a part of my final year project in University, I designed and co-developed an application that can detect potholes on roads when driven over them and mark them on a google maps interface.

A case study for the same has been written down below.

Introduction

The numerous potholes present throughout our roads cause major inconvenience to the drivers and damage to our vehicles.

Project Scope

Parkour is a pothole detection application. The main objective of this application is to avoid our users from running over potholes which could potentially damage (suspension damage, steering damage, bent or damage to rims, steering damage, flat tyres and so on) their automobiles. The application also offers navigation, pothole visuals, pothole alerts and standby notification.

Design and Implementation Constraints

Misdetection

Sensors like gyroscope and accelerometer in a cell phone can have readings due to different reasons. Even when the device is in a vehicle, there can be different reasons for the same. Narrowing down all the options, all the cell phones with the app installed, having the sensor readings at a particular location is when a pothole is being detected by the system. The limitation to this process is that the previous scenario may be due to a speed breaker as well.

Misprediction: As we are notifying the government body concerned about the pothole along with its exact location, there is a possibility that they could clear it anytime. So, if the database is not being updated frequently it could lead to is prediction.

Human Error: If the user keeps shaking their device at a particular location on purpose, then a pothole would be marked at that particular location.

R

There are several methods to demonstrate the use of accelerometer present in smartphones to detect the presence of potholes. These methods do not give accurate results as it just uses the data of the accelerometre. Hence, we include the use of gyroscope along with the accelerometer to improve the accuracy. We also deploy our models into a smartphones which makes it more applicable and practical than others.

In their study "Real time pothole detection using Android smartphones with accelerometers," Mednis et al show how cellphones can identify pothole events. Their method detects potholes with true positive rates as high as 90% using a classification scheme that flags accelerometer activity that crosses a certain z-axis threshold.

Nericell is represented as a fleet of smartphones that uses an aggregation server to assess road conditions, as well as a set of algorithms to reorient a confused smartphone accelerometer along a canonical set of axes, according to P Mohan et al.

With a mis-identification rate of 0.2 percent, Eriksson, Jakob, et al deploy a crowdsourced fleet of taxis titled Pothole Patrol to identify potholes and road irregularities using accelerometer and GPS data.

Requirements & Scope

1. The operating system in which the application is installed should be Android.
   

2. The device in which the application is installed should be a smartphone that contains accelerometer and gyroscope.
   

3. The device should be connected to a stable internet connection.
   

4. Location services should be enabled throughout the usage of the application.
   

5. While installing the application, Google Maps should be running in the background for smooth integration.

Assumptions & Dependencies

Assumptions:

Every user carries an Android phone with gyroscope and accelerometer while travelling. Every user has a stable internet connection in their phone. Google Maps API is free for up to 25,000 loads per day The Government would take necessary action once a pothole is reported.

Dependencies:

Google Maps SDK - Maps SDK for Android helps to add maps to Android apps using Google Maps data, map displays, and map gesture responses. You can also provide additional information for map locations and support user interaction by adding markers, polygons, and overlays to your map.

Google Routes APIs - The Directions API is a web service that uses an HTTP request to return JSON or XML-formatted directions between locations. You can receive directions for several modes of transportation, such as transit, driving, walking, or cycling.

Android SDK - The sensor framework, which is a part of the Android SDK, allows you to read raw data from most sensors, be it hardware or software, in an easy and consistent manner.

Overview

With the increasing population, there has been an increased load on the infrastructure of the roads. These increasing problems worsen the condition of the road because of many reasons like rains, oil spills, road accidents or inevitable wear and tear which make the road difficult to drive upon as unexpected hurdles on the road may cause more accidents and damage to the vehicle.

The potholes are detected using the accelerometer and gyroscope present in the mobile phones. When a user encounters a pothole, there is a change in the accelerometer and gyroscope readings. This deflection is noted and the location is permanently stored into the database. The previously detected potholes are indicated on the map using markers. This allows the driver to drive avoiding the pothole. Since it is an application, it can be downloaded in the user’s mobile phone free of cost.

Implementation

The result of this project is an Android Application that can detect the presence of a pothole with the help of the accelerometer and gyroscope that is present in the mobile phone.

A research was conducted by moving over more than 10 potholes and recording the values of the accelerometer and gyroscope. The average of the recorded values were taken and noted. The measurement of the gyroscope along the x axis was observed to be in the range of -1.0 rad/sec to 1.0 rad/sec, y-axis axis in the range of -0.1 rad/sec to 0.3 rad/sec and z-axis is always 0.0 rad/sec. The readings of the accelerometer along the x, y and z axes remain to be 4 m/sec2.

These detected potholes are then marked onto Google Maps.

The user downloads and opens the app “Parkour”. The user is then greeted by the welcome screen. The user had to login into the app using their username and password. If an account has not been created with the application, the user can sign up using their email id.

The login is authenticated and the map is loaded. The user is greeted by the map interface. The map would contain the potholes marked.

The pothole detected can be removed by the database administrator in case the pothole has been detected due to errors.

Conclusion

In this Case study, the accelerometer and gyroscope readings are integrated with an Android application to detect potholes that are present in the road. These readings are used to mark the potholes that are present in the roads onto the map. Thus, this application helps in the detection of potholes present in the roads and alerts vehicle drivers to evade potential accidents. The proposed approach is also economic as the application can be downloaded by the user in just a click of a button without paying an additional amount for this service.

The following system would not only help government authorities to repair the potholes present in the road but can also prevent tragic accidents and save many lives and ailing patients.

I hope you liked this case study ;). If you have any questions or queries, please do reach out to me.