# Implicit Kinematic Policies: Unifying Joint and Cartesian Action Spaces in End-to-End Robot Learning

arXiv: https://arxiv.org/abs/2203.01983 · HTML: https://mlsanity.com/paper/2203.01983/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Mar 2022.

## Built on by (most cited first, 15 of 15)

- [Visual Language Maps for Robot Navigation](https://mlsanity.com/paper/2210.05714/) (2022)
- [SE(3)-Equivariant Relational Rearrangement with Neural Descriptor Fields](https://mlsanity.com/paper/2211.09786/) (2022)
- [Diffusion Model-Augmented Behavioral Cloning](https://mlsanity.com/paper/2302.13335/) (2023)
- [Chain of Thought Imitation with Procedure Cloning](https://mlsanity.com/paper/2205.10816/) (2022)
- [On the Role of the Action Space in Robot Manipulation Learning and Sim-to-Real Transfer](https://mlsanity.com/paper/2312.03673/) (2023)
- [Where To Start? Transferring Simple Skills to Complex Environments](https://mlsanity.com/paper/2212.06111/) (2022)
- [CodeDiffuser: Attention-Enhanced Diffusion Policy via VLM-Generated Code for Instruction Ambiguity](https://mlsanity.com/paper/2506.16652/) (2025)
- [CLAS: Coordinating Multi-Robot Manipulation with Central Latent Action Spaces](https://mlsanity.com/paper/2211.15824/) (2022)
- [Multimodal Spatial Language Maps for Robot Navigation and Manipulation](https://mlsanity.com/paper/2506.06862/) (2025)
- [Redundancy-aware Action Spaces for Robot Learning](https://mlsanity.com/paper/2406.04144/) (2024)
- [SceneScore: Learning a Cost Function for Object Arrangement](https://mlsanity.com/paper/2311.08530/) (2023)
- [Learning dissection trajectories from expert surgical videos via imitation learning with equivariant diffusion](https://mlsanity.com/paper/2506.04716/) (2025)
- [Self-Adaptive VLA for Robust Robot Deployment](https://mlsanity.com/paper/2609.30092/) (2026)
- [Kinematics-Aware Diffusion Policy with Consistent 3D Observation and Action Space for Whole-Arm Robotic Manipulation](https://mlsanity.com/paper/2512.17568/) (2025)
- [Making two action heads agree: coordination mechanisms and a runtime collapse certificate for flow-matching policies](https://mlsanity.com/paper/2608.15748/) (2026)

## Abstract (arXiv)

Action representation is an important yet often overlooked aspect in end-to-end robot learning with deep networks. Choosing one action space over another (e.g. target joint positions, or Cartesian end-effector poses) can result in surprisingly stark performance differences between various downstream tasks -- and as a result, considerable research has been devoted to finding the right action space for a given application. However, in this work, we instead investigate how our models can discover and learn for themselves which action space to use. Leveraging recent work on implicit behavioral cloning, which takes both observations and actions as input, we demonstrate that it is possible to present the same action in multiple different spaces to the same policy -- allowing it to learn inductive patterns from each space. Specifically, we study the benefits of combining Cartesian and joint action spaces in the context of learning manipulation skills. To this end, we present Implicit Kinematic Policies (IKP), which incorporates the kinematic chain as a differentiable module within the deep network. Quantitative experiments across several simulated continuous control tasks -- from scooping piles of small objects, to lifting boxes with elbows, to precise block insertion with miscalibrated robots -- suggest IKP not only learns complex prehensile and non-prehensile manipulation from pixels better than baseline alternatives, but also can learn to compensate for small joint encoder offset errors. Finally, we also run qualitative experiments on a real UR5e to demonstrate the feasibility of our algorithm on a physical robotic system with real data. See https://tinyurl.com/4wz3nf86 for code and supplementary material.