Achieving autonomous navigation for biologically inspired jumping robots remains a long-standing challenge, due to the inherent instability of jumping motions and the limitations in onboard sensor capabilities. This paper proposes a skill-based hierarchical framework with dangerous action masking (SH-DAM) for autonomous navigation of jumping robot. The framework, based on hierarchical reinforcement learning, includes a low-level controller that learns locomotion skills (crawling, turning and jumping) to overcome various obstacles. A high-level controller selects and coordinates these skills, while also incorporating curriculum learning to enhance the performance of navigation tasks. For safe navigation, we utilize dangerous action masking to suppress the probability of selecting jump motions in dangerous regions. We improved the locust-inspired jumping robot platform JumpBot-S, by integrating a lightweight time-of-flight (ToF) sensor, and constructed a range of complex environments for experiments. Simulation results demonstrate that SH-DAM enables the robot to autonomously complete challenging navigation tasks. Compared to baseline algorithms, our method achieves a 12.57% increase in success rate, a 55.88% reduction in stuck rate, and a 57.89% reduction in rollover rate. Finally, we deployed our framework in real-world environments and conducted experiments in both normal lit and dimly lit conditions. This framework provides a new paradigm for jumping robot navigation in complex environments.

Locusts have various motion modes among which they continuously switch in terrestrial and aerial domains, hence achieving high environmental adaptability. Several robots have been developed to mimic the jump–gliding locomotion of locusts, but their mobility and transitional stability are limited because of structural and control limitations at a small scale. In this article, we develop a small-scale locust-inspired robot (LocustBot) that can not only jump and glide but also crawl. We propose a coordinately actuated mechanism that allows LocustBot to perform jump–gliding with few actuators. To achieve the stable and long-distance moving, a reinforcement-learning-based optimized control is used to generate then track the robot's position and orientation from take-off to landing. The jump–gliding distance of LocustBot reaches 5.39 m, revealing a high-energy utilization efficiency of the mobile strategy, which combines the spring-driven jumping with the propeller-driven gliding. Remarkably, without a high platform, the robot can still achieve a far moving range by continuous crawl–jump–gliding on horizontal planes and, thus, outperforms the state-of-art jump–gliding robots.