Abstract
End-to-end autonomous driving has achieved remarkable progress by integratingperception, prediction, and planning into a fully differentiable framework.Yet, to fully realize its potential, an effective online trajectory evaluationis indispensable to ensure safety. By forecasting the future outcomes of agiven trajectory, trajectory evaluation becomes much more effective. This goalcan be achieved by employing a world model to capture environmental dynamicsand predict future states. Therefore, we propose an end-to-end drivingframework WoTE, which leverages a BEV World model to predict future BEV statesfor Trajectory Evaluation. The proposed BEV world model is latency-efficientcompared to image-level world models and can be seamlessly supervised usingoff-the-shelf BEV-space traffic simulators. We validate our framework on boththe NAVSIM benchmark and the closed-loop Bench2Drive benchmark based on theCARLA simulator, achieving state-of-the-art performance. Code is released athttps://github.com/liyingyanUCAS/WoTE.