Sequential Planning via Anchored Robotic Keypoints
2026-06-29 • Robotics
Robotics
AI summaryⓘ
The authors introduce SPARK, a system that helps robots plan and perform tasks without needing training. It works by creating a plan using predefined robot actions and focuses on improving how the robot recognizes objects, which is where most mistakes happen. SPARK uses multiple attempts to find the best way to identify objects and can retry actions without starting over, improving success rates across different robots and tasks. This approach allows different parts of the system to be upgraded independently and provides useful data for future learning methods without needing manual guidance.
neurosymbolic manipulationbehavior treerobotic keypointslarge language modelsperceptiontask planningzero-shot learningrecovery looprobot controlobject detection
Authors
Bryce Grant, Aryeh Rothenberg, Logan Senning, Zonghe Chua, Zach Patterson, Peng Wang
Abstract
We present Sequential Planning via Anchored Robotic Keypoints, SPARK, a training-free neurosymbolic manipulation system that reaches 43.7% on six LIBERO-PRO position \& task cells, more than doubling CaP-Agent0 and Vision-Language-Action (VLA) baselines. CaP-Agent0, a multi-turn code-generation agent, achieves 18.2% by re-querying an LLM at every turn, but its restart-from-scratch solution proves costly against minor policy failures. Perception is the layer that fails most under position and task changes so SPARK spends its computation there. A single Gemini call composes the plan as a typed behavior tree (BT) of composable primitives, each already containing the low-level control (motion, grasping, depth geometry) a code-generation agent would otherwise regenerate on every trial. The rest of the budget goes to perception: a second Gemini call proposes three alternative text prompts per object, SAM3 evaluates each, and we keep the prompt$\to$label pair with the most confident detection and a recovery loop then retries a failed primitive against freshly detected objects, with no new LLM call. The alternative prompts add +27.7 points on the spatial suite and +10.0 on the object suite, with the recovery loop adding +5.0 overall. SPARK runs the same primitives on three robot families (UR10e, Franka FR3, bimanual Franka) across nine unique tasks at twenty trials each, averaging 68%. Since the detector, planner, and controller modules sit behind the typed plan, they swap independently without training, and each primitive's checkable post-condition traces a failure to the corresponding module or a kinematic limit. Every trial logs a verified, labeled trajectory, so a training-free planner that already beats VLAs can supply the data those policies need without teleoperation. Project page: https://cwru-aism.github.io/spark-page/