Classify Object Behaviour to Enhance the Safety of Autonomous Vehicles

This paper presents a MATLAB-based simulation model for autonomous vehicle behavior in dynamic urban environments, emphasizing real-time decision-making and environmental awareness. The core objective of the simulation is to replicate intelligent vehicle responses to various real-world scenarios including pedestrian interaction, traffic signals, inter-vehicle navigation, and environmental challenges like fog. The vehicle is programmed to estimate the speed of a crossing pedestrian and respond appropriately to ensure safety by stopping when necessary. It adheres to traffic signal rules by halting at red lights and resuming motion upon green. In-lane obstacle detection is continuously performed to monitor the presence of slower vehicles. When an obstacle is detected in the current lane, the simulation evaluates the adjacent lane for safety. If the adjacent lane is clear, the vehicle initiates a lane change with a visual turn signal. In cases where the adjacent lane is also obstructed, the vehicle slows down and, after a delay, honks to alert nearby drivers. The simulation also integrates a fog effect to test reduced visibility scenarios, under which the vehicle automatically adjusts its speed. Position, lane, and velocity updates are computed in real time based on sensory inputs and environmental feedback. A graphical interface displays the animated road, dynamic lane lines, traffic lights, other vehicles, pedestrians, and the vehicle’s maneuvers, allowing users to visualize the full autonomous driving scenario. This model provides a foundational platform for testing and developing more advanced algorithms in autonomous driving and environment-based decision systems.

Index Terms— Autonomous vehicle, MATLAB simulation, pedestrian detection, traffic signal recognition, lane change, obstacle avoidance, fog simulation, real-time decision-making, vehicle speed control, intelligent transportation systems.