In this video we demonstrate the performance of a reach-avoid controller designed via approximate dynamic programming for a nonlinear tire-force car model with 6 states and 2 inputs. The task is to avoid the obstacle (static car) while reaching the end of the turn (denoted by a dashed line). A simple control strategy is used to drive the car to an initial state (position, orientation, velocity) from which the reach-avoid controller takes over . We carried out two types of experiments, one with a linear velocity of 0.5m/s and one with a linear velocity of 2m/s. More information on the design approach can be found in [ Ссылка ].

свернуть