Legged robots have demonstrated remarkable potential for dynamic locomotion and terrain adaptability, making them a prominent focus of research. However, achieving robust and agile bipedal running remains challenging due to the complex dynamics of legged locomotion. In this paper, we propose a reinforcement learning framework for robust bipedal running, incorporating a simple reference trajectory generator and an asymmetric actor-critic architecture. The reference generator, based on kinematics, provides diverse trajectory references while preserving key gait characteristics, facilitating efficient policy exploration. To mitigate the simulation-to-reality gap, we extract latent variables encoding environmental and motion information from dual historical observations. Our method simplifies the trajectory generation process while maintaining effective guidance for learning. Extensive simulation and physical experiments demonstrate that, compared to model-based and learning-based baselines, our approach achieves higher agility, more accurate velocity tracking, and stronger disturbance rejection while preserving gait stability. The resulting controller exhibits spring–mass running dynamics that remain robust on both flat and uneven terrains.

Hexapod robots are well suited for disaster rescuing tasks due to their stability and load capability. However, most current hexapod robots still rely on static gaits that largely limit their locomotion speed. This paper introduces a hierarchical control strategy to realize a dynamic alternating tripod trotting gait for a hexapod robot based on multi-modal impedance control. At the low level, a position-based impedance controller is developed to realize an adjustable compliant behavior for each leg. At the high level, a new gait controller is developed to generate a stable alternating tripod trotting gait, in which a gait state machine, a leg compliance modulation strategy, and a close-looped body attitude stabilizer are imposed. As a result, the alternating tripod trotting of the hexapod robot can be synchronized as the running of a bipedal robot with stable body attitude. Moreover, this control strategy was verified by experiments on a newly designed pony-sized disaster rescuing robot, HexbotIV, which successfully achieved a dynamic trotting gait with ability to resist the disturbances of mildly uneven terrains. Our control strategy as well as the experimental study can be a valuable reference for other hexapod robots and thus paves a way to the practical deployment of disaster rescuing robots.