Robots navigating in a robot–human coexisting environment need to optimize their paths not only for task-related performance (e.g. safety and efficiency) but also for their social compliance to other pedestrians. This is a crucial yet challenging task. Previous work has shown the power of deep reinforcement learning (DRL) techniques by employing them to train efficient policies for robot navigation. However, their performance deteriorates when the crowd size grows. We cope with this problem by allowing the robot to keep an adapting distance from the pedestrians and perform safe navigation even in high density environments. We first derive a quantitative formula representing the relationship between uncomfortable distance and pedestrian density from a real-word tracking dataset. Then this formula is applied in reward shaping of DRL to get resilient reward functions (R2F). Qualitative and quantitative evaluation results demonstrate that our method outperforms state-of-the-art methods in both low and high pedestrian density environments. GRAPHICAL ABSTRACT

Mobile robots operating in public environments require the ability to navigate among humans and obstacles in a socially compliant and safe manner. Previous work has shown the power of deep reinforcement learning (DRL) techniques by employing them to train efficient policies for robot navigation. However, most DRL-based robot navigation methods only consider dynamic pedestrians and do not take static obstacles into account. The ability to approach pedestrians and obstacles differently will improve a robot's navigation efficiency. In this work, we propose a novel network, the obstacle-robot uni-action (ORU) network, to encode the one-way direct effects of obstacles on the robot. Obstacles' indirect effects on the robot, represented by the obstacle-human uni-action (OHU), together with human–human interaction (HHI), are concatenated to a human–robot interaction (HRI) network to obtain the features of a crowd's effects on the robot. We also implement a variable reaching goal reward and an approaching goal reward in our model, which can enhance the method's performance in terms of navigation time. Results of test experiments in both simulation and real datasets demonstrate that the proposed method outperforms state-of-the-art methods. GRAPHICAL ABSTRACT

Active acoustic signaling can improve robot navigation in crowds, but methods based on a cooperative-response assumption may generate unnecessary signals near static pedestrians and misjudge signaling effectiveness when non-cooperative pedestrians approach. This study proposes heterogeneous interactive crowd navigation with acoustic signaling, a selective acoustic-signaling method for heterogeneous crowds. The proposed method models dynamic cooperative pedestrians, static non-responsive pedestrians, and approaching non-cooperative chasers. A differentiated reinforcement-learning social reward suppresses ineffective signaling in static-only neighborhoods, while a second training stage introduces non-cooperative chasers controlled by a standard optimal reciprocal collision avoidance policy with fixed parameters to improve motion avoidance when signaling cannot alter chaser motion. Experiments in corridor and square-hall environments used 500 test episodes per setting. The differentiated reward reduced total signaling rates by 41.2% and 47.0%, respectively, and reduced static-condition signaling rates to 0%, while largely preserving dynamic-condition signaling and navigation time. In complete heterogeneous scenarios, the proposed method achieved success rates of 0.94 and 0.91, collision rates of 0.02 and 0.04, and catch rates of 0.08 and 0.12. These results indicate improved signaling selectivity and navigation performance under the evaluated controlled simulation conditions involving predefined cooperative and non-responsive pedestrian behaviors.