Route Planning and Machine Learning Algorithms for Sensor-Equipped Autonomous Vehicles in Agriculture

Mihai PETRESCU, Maria-Cristina ȘTEFAN, Andreea Anamaria PANAGOREȚ, Ioana Octavia CROITORU, Sorin CHIRIȚĂ, Bogdan Cătălin STERIOPOL

Abstract

This paper aims to improve the efficiency and safety of autonomous agricultural vehicles (AAVs) in complex agricultural settings by integrating sensor technology and deep reinforcement learning techniques. The traditional fixed-route transport is expensive and prone to vehicle collisions. Advanced AAVs equipped with multiple sensors were employed for collecting and processing sensor data, were employed for collecting and processing sensor data. This algorithm can effectively identify moving AAVs, static obstacles and other relevant targets in agricultural environments. A deep reinforcement learning model was built by using Deep Q-Networks and neural networks and a simulation environment was created with the purpose of validating the path planning and obstacle avoidance capabilities of the proposed model.

Keywords

Simulation, Kinematics, Dynamics, Trajectory, Cornering.

View full article