Autonomous Quantum Reinforcement Learning for Robot Navigation

Sie sind hier: Homepage > Suche

Autonomous Quantum Reinforcement Learning for Robot Navigation

Mohan, Arjun / Jayabalan, Sudharsan / Mohan, Archana

Abstract In order to achieve a safe and traffic free transportation from place to place it is mandatory that a vehicle should communicate with other vehicles autonomously to control their speed and movement. Fuzzy logic control architecture which is placed in the vehicle has been designed to take appropriate decision based on the various parameters inside the vehicle. Number of input and output parameters and rules of operation can be varied easily in the fuzzy logic toolbox in MATLAB. The output will be sent to the micro-controller using RS232 serial communication. Now the micro-controller will take the appropriate action to execute the command received. So in order to communicate with other vehicle we have used GPS receiver which will transmit the current position of the vehicle to nearby vehicle ranging within 30 or 100 m using Zigbee communication in a common transmission frequency. If there is any vehicles present in this range it will accept and communicate back by sending the corresponding GPS location of that vehicle and the path of its course. If it is found to be moving on the same direction or path a communication with the vehicle will be established and data will be transmitted among them to achieve the safe and smooth traffic travel to increase the vehicle efficiency. The same rules will apply if there is more than one vehicle is present in the transmitting range. All these transmitting data will be privacy protected. We have successfully controlled prototype model and found that architecture is working autonomously up to the expectation and provides efficient travel for the vehicle.