PAL Robotics · Humanoid · 2023
Kangaroo, from Barcelona's PAL Robotics, is a bipedal humanoid robot purpose-built for research on agile and dynamic locomotion. Its signature engineering is mechanical: instead of conventional rotary joints, Kangaroo uses custom linear actuators with integrated force sensors arranged in closed/parallel kinematic chains, concentrating mass away from the legs to achieve very low moving inertia — which translates into high joint speeds and torques, better impact resilience and reduced energy consumption. It features 12 actuated degrees of freedom per leg and has demonstrated real-hardware jumping, showcasing the platform's robustness and resilience. PAL positions Kangaroo as a testbed for dynamic locomotion, reinforcement learning and embodied AI, complementing its established humanoid and mobile robots (TALOS, TIAGo, REEM, ARI). NOTE: leg DOF and actuation from PAL's Kangaroo hardware paper (12 DOF/leg, linear actuators); exact height/weight/full-body DOF vary by configuration and are flagged for outreach rather than guessed.
Price on application
View full interactive profile, comparisons & videos →| Category | Humanoid |
| Sub-type | Agile Bipedal Locomotion Research Platform |
| Status | Active |
| Year | 2023 |
| Origin | Spain |
| Announced | 2023 |
| Released | 2023 |
| Weight | 40 kg |
| Degrees of freedom | 28 |
| Payload | 6 kg |
| Battery life | 180 min |
| AI model | Closed-loop force/torque controllers at joint level; whole-body planning/control pipeline for jumping/dynamic locomotion; optional NN/RL via Jetson |
| Operating system | ROS-based |
| Actuator type | Custom force-controlled ball-screw linear actuators with integrated force sensors; closed/parallel kinematic chains; decoupled leg-length from hip-swing; 1-2 kHz EtherCAT control |
| Use cases | Dynamic-locomotion research, Reinforcement learning, Embodied-AI research, Legged-robot control R&D |
| Made in | Spain (Barcelona) |
