Sensor based object tracking and collision warning/avoidance system
Ultrasonic and infrared sensors have critical importance due to low cost and easy to control. They require low computations. In the automobile sector, collision warning and avoidance is becoming a critical issue. Therefore, sensors play an important role to detect the objects, alert the driver, and for automatic collision avoidance. Likewise, collision avoidance is essential for robots to move properly in the uncertain environment. In industrial and other applications, robot tracking is required to move the robot autonomously. Thus, collision avoidance and tracking systems should be improved to achieve high efficiency.
Sensor-based Object Tracking System
We are using two approaches to implement the system using simple binary logic controller and fuzzy logic controller. The flow chart of the object tracking system is shown in the following Figure. Initially, the sensors determine the position of the object, and then, different operations are performed on the range data to obtain the final decision. Finally, commands are sent to the DC and servo motors.
To follow an object, the robot requires range measurements from the front, left, and right directions. Therefore, we use five range sensors for the tracking system. The STM sensor is used for the forward and backward motions, and the remaining sensors aree used for the left and right rotations. The STL, STM, and STR sensors give the range measurements from the front side. The SL sensor gives the range information from the left side; likewise, the SR sensor gives the range measurements from the right side. The STM sensor is reserved for the forward and backward motions, and the other four sensors aree used for the left and right rotations. Fifteen different obstacle environments used for object tracking are shown in the following FIgure.
Real time video of object tracking robot
Sensor-based Collision-warnning Avoidance System (CWAS)
The idea behind the system is that the control unit receives information from various sources (sensors and remotely controlled device) and performs specific tasks. The control unit contains a microcontroller that sends commands to all modules (motor driver and alarms). To avoid collision, information on the surroundings should be known from all sides (left, right, front, and back). For this purpose, eight range sensors are that provide range information. Three sensors are placed on the front side of the robot. Similarly, three sensors are placed on the back side, and the other sensors are mounted on the left and right sides. The arrangement of sensors on the robot is shown in the following Figure.
A remotely controlled device is employed to move the robot manually. DC and servo motors are used to move the robot. The DC motor was used for the forward and backward motions. The servo motor was used for the left and right motions. The overview of the CWAS is illustrated in the following Figure.
1. Integrated tracking and collision warning/avoidance system
Recently, I am working on this project. I am using two approaches to implement the system using simple binary logic microcontroller and fuzzy logic controller.
The proposed approach relied on techniques in which the positions of static and moving objects were determined. The system was tested with a mobile robot. It was controlled through instructions by human and through the range sensors. The robot can be used in two ways. First, for tracking an object autonomously by maintaining a constant distance from it. It can also be used for collision prediction, which has a critical importance in terms of alerting the driver before an accident, and collision avoidance in which a safe distance is predefined for each function; thus, both are combined in order to increase the safety level.
In the object tracking system, the robot was autonomously instructed regarding the navigation. We conducted experiments for various scenarios. The robot successfully tracked the object by maintaining a constant distance of 30 cm from the tracked object in various situations.
In the CWAS, the robot performed security functions precisely. It is concluded that collision prediction, achieved via the range measurements, helps to reduce vehicle accidents and that the driver can be notified before an accident via light and sound alarms.
The conventional binary logic controller and the FLC were utilized in order to handle the uncertain data from the sensors. Both controllers performed well for both approaches. Since the FLC can handle uncertainties well, smoothness was achieved in terms of object tracking. It demonstrated precise results in terms of object tracking. However, computational complexity was high in the fuzzy approach.
Future work will focus on creating a well-defined sensor architecture by increasing the sensor precision. In object tracking, the speed variation of the robot will be improved by estimating the speed of the followed vehicle precisely, and a separate control unit will be installed in order to identify the tracked object.