This paper presents a novel design of control algorithms for lane change assistance and autonomous driving on highways, based on recent results in Scenario Model Predictive Control (SCMPC). The basic idea is to account for the uncertainty in the traffic environment by a small number of future scenarios, which is intuitive and computationally efficient. These scenarios can be generated by any model-based or data-based approach. The paper discusses the SCMPC design procedure, which is simple and can be generalized to other control challenges in automated driving, as well as the controller's robustness properties. Experimental results demonstrate the effectiveness of the SCMPC algorithm and its performance in lane change situations on highways.
We propose a deep reinforcement learning approach for solving a mapless navigation problem in warehouse scenarios. In our approach, an automatic guided vehicle is equipped with two LiDAR sensors and one frontal RGB camera and learns to perform a targeted navigation task. The challenges reside in the sparseness of positive samples for learning, multi-modal sensor perception with partial observability, the demand for accurate steering maneuvers together with long training cycles. To address these points, we propose NavACL-Q as an automatic curriculum learning method in combination with a distributed version of the soft actor-critic algorithm. The performance of the learning algorithm is evaluated exhaustively in a different warehouse environment to validate both robustness and generalizability of the learned policy. Results in NVIDIA Isaac Sim demonstrates that our trained agent significantly outperforms the map-based navigation pipeline provided by NVIDIA Isaac Sim with an increased agent-goal distance of 3 m and a wider initial relative agent-goal rotation of approximately . The ablation studies also suggest that NavACL-Q greatly facilitates the whole learning process with a performance gain of roughly compared to training with random starts and a pre-trained feature extractor manifestly boosts the performance by approximately .
We propose a 2D path planning algorithm in a non-convex workspace defined as a sequence of connected convex polytopes. The reference path is parameterized as a B-spline curve, which is guaranteed to entirely remain within the workspace by exploiting the local convexity property and by formulating linear constraints on the control points of the B-spline. The novelties of the paper lie in the use of the equivalent Bézier representation of the B-spline curve, which significantly reduces the conservatism in the local convexity bound and in the integration of these constraints into a convex quadratic optimization problem, which minimizes the curve length. The algorithm is successfully validated in both simulations and experiments, by providing obstacle-free reference paths on real occupancy grid maps obtained from the laser scan data of a mobile robot platform.
To deal with the problem of optimal path planning in 2D space, this paper introduces a new toolbox named “Navigation with Polytopes” and explains the algorithms behind it. The toolbox allows one to create a polytopic map from a standard grid map, search for an optimal corridor, and plan a safe B-spline reference path used for mobile robot navigation. Specifically, the B-spline path is converted into its equivalent Bézier representation via a novel calculation method in order to reduce the conservativeness of the constrained path planning problem. The conversion can handle the differences between the curve intervals and allows for efficient computation. Furthermore, two different constraint formulations used for enforcing a B-spline path to stay within the sequence of connected polytopes are proposed, one with a guaranteed solution. The toolbox was extensively validated through simulations and experiments.
25 September 2024 in IEEE Transactions on Control Systems Technology
This article addresses the problem of traffic prediction and control of autonomous vehicles on highways. An interacting multiple model Kalman filter (IMM-KF)-related algorithm is applied to predict the motion behavior of the traffic participants by considering their interactions. A scenario generation component is used to produce plausible scenarios of the vehicles. A novel integrated decision-making and control system is proposed by applying a scenario-based model predictive control (MPC) approach. The designed controller considers safety, driving comfort, and traffic rules. The recursive feasibility of the controller is guaranteed under the inclusion of the “worst case” as an additional scenario to obtain safe inputs. Finally, the proposed scheme is evaluated under a high-fidelity IPG CarMaker and Simulink co-simulation environment. Simulation results indicate that the vehicle performs safe maneuvers under the designed control framework in different traffic situations.