| Sim2Real | Simulation-to-Real transfer, the core paradigm | throughout |
| BAM | Bandwidth-Aware Motor, a bandwidth-aware servo actuator model | BAM Actuator Modeling and Identification |
| MJCF | MuJoCo robot-modeling file format | Automated Onshape→MJCF Modeling |
| PPO | Proximal Policy Optimization algorithm | PPO Multi-Task Training |
| Domain Randomization | randomly perturbing simulation parameters during training | Domain Randomization |
| Curriculum Learning | progressing training from easy to hard | Domain Randomization |
| ONNX | Open Neural Network Exchange, a model-interchange format | PPO Multi-Task Training, 14 |
| MuJoCo-Warp | NVIDIA GPU-parallel physics simulation | The MuJoCo-Warp Simulation Environment |
| mjlab | Pollen's in-house robotics simulation library | The MuJoCo-Warp Simulation Environment |
| TTL bus | half-duplex serial communication bus (used by servos) | Hardware Circuits and the TTL Bus |
| Backlash | hysteresis caused by gear clearance | BAM Actuator Modeling and Identification |
| Randomization intensity | the range of the randomly perturbed parameters | Domain Randomization |
| Sim2Real gap | the dynamics gap between simulation and reality | Simulation Debugging and Diagnostics, 15 |
| System identification | inversely identifying true parameters | The Sim2Real Closed-Loop Iteration |
| RKNN | RK-platform neural-network inference toolkit | RK3566 Image and Firmware, 14 |
| BMS | Battery Management System | Hardware Circuits and the TTL Bus |
| onshape-to-robot | CAD→MJCF automatic export tool | Automated Onshape→MJCF Modeling |
| Observation space | the sensor-data vector fed to the policy | The MuJoCo-Warp Simulation Environment, 12 |
| Action space | the control-command vector output by the policy | The MuJoCo-Warp Simulation Environment, 12 |