MultiVehicle Simulator (MVSim): Lightweight dynamics simulator for multiagents and mobile robotics research

Development of applications related to closed-loop control requires either testing on the field or a realistic simulator, with latter being more convenient, inexpensive, safe, and leading shorter development cycles. To address that need, present work introduces MVSim, simulator for multiple vehicles robots capable running dozens agents in simple scenarios, handful them complex scenarios. MVSim employs physics-grounded friction models tire-ground interaction, aims at accurate GPU-accelerated simulation most common modern sensors employed mobile robotics autonomous vehicle research, such as depth RGB cameras, 2D 3D LiDAR scanners. All depth-related are able accurately measure distances provided by user define custom world elements. Efficient is achieved means focusing ground vehicles, which allows use simplified physics engine body collisions while solving wheel-ground interaction forces separately. The core parts system written C++ maximum efficiency, Python, ROS 1, 2 wrappers also offered easy integration into systems. A publish/subscribe protocol based ZeroMQ (ZMQ) defined allow multiprocess access modify simulation. This enables makes easier do research vehicular dynamics, navigation algorithms, simultaneous localization mapping (SLAM) methods.