Automatization of a Robotic Electric Vehicle with Electronic Differential

Abstract

A large amount of the scientific research in Mobile Robotics in the last years has focused in the development of autonomous urban vehicles, which supposes navigation in a structured environment like flat asphalt, allowing the knowledge of important variables and also simplifications with respect to the tire-ground force interactions. However, for the case of the control of all-terrain vehicles, like tractors, the knowledge of the adherence properties like slipping, skidding, trepidations and high slopes is very difficult and constitutes a great challenge for the scientific and technological research all over the world. One of these challenges is the approach of the distribution of torque/speed in the non-steerable wheels of a multitraction vehicle, known as electronic differential distribution, a topic that was addressed in Project VERDE - Robotic Electric Vehicle with Electronic Differential (FAPESP regular n. 2014 / 02672-9), from the same research group. This proposal presents itself as a continuation of the previous VERDE project where two mini-vehicles with electric motorization were built (with two independent rear motors), instruments with inertial navigation system INS / GPS, on-board computer and CAN bus, Laser, camera and sensors on wheels and motors.The objectives of this new proposal are threefold: (1) Technological improvement and validation of the sensory assembly of the all-terrain mini-vehicles of the original VERDE project, including the insertion of suspension deflection sensors; (2) Experimental development, implementation and validation of linear and nonlinear trajectory control techniques for application in terrains with extreme conditions of adhesion, slipping and skidding, making use of the electronic differential already available in the vehicle. The use / evolution of the torque / speed distribution strategies in the independent wheels of the multitraction vehicle (Cordeiro, 2017b), (Ribeiro, 2016), developed in the original VERDE project, and also in phase of improvement in the PhD thesis of Ribeiro (2016); (3) Experimental deployment, implementation and validation of sensory perception techniques such as SLAM (visual and laser) and autonomous convoy using computer vision.Thus, using the great potential of the scaled vehicles available from the original VERDE project, we believe that the Auto_VERDE project will allow the evolution and experimental validation of the control approaches developed and / or under development for applications in vehicles for extreme conditions terrains like those used in agricultural applications. (AU)