Hello everyone,
I’m trying to train an RL agent on a vehicle to brake when encountering a pedestrian. I have set up a scenario using the driving scenario toolbox and everything in simulink to make it work.
I just have one problem: I’d like to randomize the pedestrian’s position, so I created a random function in the scenario definition script, but when I use fast restart the scenario reader gets stuck on the first simulation scenario. If I don’t, Simulink recompiles and everything works.
Is there a way to not disable fast restart, but still completely restart the scenario with variations?Hello everyone,
I’m trying to train an RL agent on a vehicle to brake when encountering a pedestrian. I have set up a scenario using the driving scenario toolbox and everything in simulink to make it work.
I just have one problem: I’d like to randomize the pedestrian’s position, so I created a random function in the scenario definition script, but when I use fast restart the scenario reader gets stuck on the first simulation scenario. If I don’t, Simulink recompiles and everything works.
Is there a way to not disable fast restart, but still completely restart the scenario with variations? Hello everyone,
I’m trying to train an RL agent on a vehicle to brake when encountering a pedestrian. I have set up a scenario using the driving scenario toolbox and everything in simulink to make it work.
I just have one problem: I’d like to randomize the pedestrian’s position, so I created a random function in the scenario definition script, but when I use fast restart the scenario reader gets stuck on the first simulation scenario. If I don’t, Simulink recompiles and everything works.
Is there a way to not disable fast restart, but still completely restart the scenario with variations? fast restart, driving scenario, reinforcement learning toolbox, reset scenario MATLAB Answers — New Questions

​