Real Time Distance and Speed Evaluation Using Mobile Robot

Author Name(s): Rahul Gupta, Ramchand Hablani, Susmita Guin
Author Email: rahul37g@gmail.com

Abstract

We all are aware that the future is of artificial intelligence and we want that machine should do smart work. Sometimes, due to hard work driver is not able to concentrate properly while driving the vehicle, also if driver is driving the vehicle continuously for a long time, he may suffer from fatigue in eyes and may not be able to focus on road properly, which may lead to accident if driver fails to detect the presence of obstacle. To overcome this problem, we need an automated vehicle, which can detect the obstacle and can calculate the distance along with the current relative speed of the vehicle and can perform the action accordingly.

In this paper, we introduce the Real Time Distance and Speed Evaluation (RTDSE) system using mobile robot which detects the presence of object before accident and calculate the distance between the robot and obstacle along with the current relative speed of the robot, which helps system to perform actions like reducing the speed of robot, overtaking the object or to apply emergency brakes.

Keywords

Obstacle Detection, Distance Measurement, Robot, Speed Measurement, Stereo Vision, ZED Camera.

Introduction

Due to increasing challenges in day to day life and to reach the destination on time, time plays an important factor. To save time, people drive vehicle fast on road and as a result, it leads to an accident. One of the main reasons is human error, which can be reduced but cannot be avoided completely. A human error like fatigue in eyes due to whole day hard work or by driving nonstop for a long time which makes driver exhaust and as a result, he fails to detect the obstacle or vehicle on the road which leads to an accident. If the driver is drunk, he may lose control over the vehicle and can cause damage to the vehicle or in the worst-case scenario may lead to a dangerous accident and as a result can cause his or others life at great risk.

Another reason is unexpected obstacles like an animal, pedestrian or another vehicle which suddenly appears in front of the vehicle and as a result driver gets dazed and fails to apply emergency brakes or lose control over vehicle due to the sudden appearance of obstacles.

To minimize this human error which is causing an accident, an automated system should be used which act as a driver assistance system for controlling the speed of the vehicle or to apply emergency brakes. To achieve the above-mentioned goals, we are proposing a Real Time Distance and Speed Evaluation (RTDSE) system. This system will detect obstacles in front of the robot and avoid it, with short response time. Our aim is to eliminate the cause of accident produced by human error by introducing RTDSE using the mobile robot. The system collects information from various types of sensors mounted on the robot and performs computation to detect if any obstacle is present. If the difference between the speeds of both robot and object is significant then the system will apply the brakes or bypass the obstacle.

Conclusion

In this paper, we presented a real-time, distance and speed evaluation using the mobile robot system, to detect the object by using stereo vision method with the help of ZED camera mounted on the robot. Further, we have also calculated the distance of the moving object along with the relative speed.

A new framework was developed for the RTDSE which makes the use of stereo camera, ultrasonic and IR range sensors to get the information about the position of objects and relative speed of the vehicle. Testing of the proposed algorithm on higher vehicle speed by using the stereo camera or by employing long-range sensors and hardware components having higher computing speed and faster response time has also been done. The proposed method is able to take appropriate control action if it finds any static or moving object. RTDSE was able to avoid a collision in all cases tested in the real environment. However, there is still a room for improvement. Our future work includes improvement of the system performance by applying Machine learning methods such as supervised learning algorithms. It will not only improve the efficiency of the system but also simplify the work by detecting the given object accurately. We will try to develop a classifier which will take useful features to detect an object from its background.

1,236 total views, no views today

Download PDF File

About the author: admin