Machine Learning Based Autonomous Road Maintenance System Using Cold Lay Asphalt

Author Name(s): Abhinav Sreesan, Anirudh Shankar, Vignesh Vaidyanathan, Shubhangi Kharche
Author Email: abhinav.abhinav@siesgst.ac.in

Abstract

Road maintenance is quotidian problem faced by the concerned authorities as it requires time, resources and labor and if not treated with utmost alacrity, may result in mishaps and accidents. The following paper suggests an innovative approach that can reduce the time and labor requirement by incorporating concepts of image processing and machine learning with cutting edge materials like Cold Lay Asphalt to give an efficient solution to the afore mentioned problem. The primary goal of this paper is to identify and highlight an application of autonomous vehicles. To this end, the following paper proposes a machine learning based autonomous road maintenance system (ML-ARMS). The vehicle (Bot) in ML-ARMS is trained and controlled using Raspberry-Pi. The accuracy of the Machine Learning algorithm for 500 training images is found to be 42.778%.

Keywords

Autonomous Vehicle, Cold Lay Asphalt, Road Maintenance.

Introduction

Road maintenance has been a tedious task especially in a country like India with over hundred thousand kilometres of roadways that have to be constantly maintained. Every year a major part of the Government’s budget goes towards the maintenance of these roads. Thus, with advancements in vehicle automation technology it becomes possible to automate the process and to have a completely independent system that can be monitored remotely; also new road surface materials have made it possible for a quick and efficient maintenance process. This reduces a boatload of money spent in road maintenance and also provides scores of employment opportunities in the field of science and technology. The proposed system used for the maintenance of the damaged/faulty roads has sensors fitted on-board the vehicle which accurately detects the aberrations on the surface and then provide remedies by cleaning it and resurfacing the area to provide a smooth and clean surface for traversing. This will be extremely beneficial to avoid road accidents and won’t proliferate traffic in that area. Now, for the development of this model ultrasonic distance sensors and cameras have been used to accomplish the proposed model. A vehicle is built which runs autonomously with the assistance of camera and two sensors to detect obstacles, road signs & signals. Sensors are attached to the bottom of the vehicles to detect abnormalities of the road so as to fix it. A special kind of road surfacing material called cold lay asphalt has been developed recently which can be used to cure the roads and make sure that the roads are smooth enough for traversing. The cold lay asphalt is a type of mixture analogous to that of concrete as it doesn’t need any type of heating and is used at variety of temperature range. The material is a lot like conventional asphalt as it takes the shape of the potholes just after dispensing the material.

Conclusion

This proposition is a completely autonomous vehicle capable of traversing the streets; it can detect instructions on the road and traffic lights using image processing and can detect hurdles in its path using an array of ultrasonic sensors and react accordingly. Aberrations on the tarmac are detected using an array of IR Photodiodes and then the Asphalt Material is dispensed on the area thus making the road smooth and pothole free. With progress and innovations in this concerned field, the price incurred on such a vehicle can be brought down which will enable them to be widely used. The system can then be upgraded to a mesh of connected vehicles that can communicate in real time to improve speed and efficiency of the task. Machine learning algorithms have increased the proficiency of such cars. More advanced technology like LIDAR can be implemented to enable the vehicle to traverse in places with rougher terrain.

The system with small modifications can be implemented in a variety of other application so as to automate daily routines like an autonomous garbage collection vehicle or an autonomous road laying vehicle. It can also be used in the industrial sector to perform tasks in hazardous environments. The ML-ARMS if adopted on a wider scale will benefit the society and will render accidents due to road conditions obsolete. Further developments can be made to the proposed system to implement a completely autonomous road laying and monitoring system that can be implemented in laying roads in remote and rural areas without the need for any human intervention.

1,303 total views, no views today

Download PDF File

About the author: admin