cs.ROMar 19, 2026

Robotic Agentic Platform for Intelligent Electric Vehicle Disassembly

Authors: Zachary Allen, Max Conway, Lyle Antieau, Allen Ponraj, Nikolaus Correll

Organizations: Department of Computer Science, University of Colorado, Boulder

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Electric vehicles (EV) create an urgent need for scalable battery recycling, yet disassembly of EV battery packs remains largely manual due to high design variability. We present our Robotic Agentic Platform for Intelligent Disassembly (RAPID), designed to investigate perception-driven manipulation, flexible automation, and AI-assisted robot programming in realistic recycling scenarios. The system integrates a gantry-mounted industrial manipulator, RGB-D perception, and an automated nut-running tool for fastener removal on a full-scale EV battery pack. An open-vocabulary object detection pipeline achieves 0.9757 mAP50, enabling reliable identification of screws, nuts, busbars, and other components. We experimentally evaluate (n=204) three one-shot fastener removal strategies: taught-in poses (97% success rate, 24 min duration), one-shot vision execution (57%, 29 min), and visual servoing (83%, 36 min), comparing success rate and disassembly time for the battery's top cover fasteners. To support flexible interaction, we introduce agentic AI specifications for robotic disassembly tasks, allowing LLM agents to translate high-level instructions into robot actions through structured tool interfaces and ROS services. We evaluate SmolAgents with GPT-4o-mini and Qwen 3.5 9B/4B on edge hardware. Tool-based interfaces achieve 100% task completion, while automatic ROS service discovery shows 43.3% failure rates, highlighting the need for structured robot APIs for reliable LLM-driven control. This open-source platform enables systematic investigation of human-robot collaboration, agentic robot programming, and increasingly autonomous disassembly workflows, providing a practical foundation for research toward scalable robotic battery recycling.

Figures & tables

Fig. 1 : Hyundai Ioniq5 EV battery being dismantled by the RAPID system
Fig. 2 : Integration of conventional automation systems, open-world vision systems, motion planning and Agentic AI.
TaskDescriptionTime (s)WorkersMan-secondsTools
1Remove 8 large screws on top cover1202240Large nutrunner
2Remove 70 screws and nuts holding top cover50121002Nutrunner (high torque)
3Remove cover plate27254None
4Remove two busbars to break battery voltage into two1002200Nutrunner (high torque)
5Remove two large cables connecting across battery1732346Nutrunner (high torque)
6Remove 6+1 large busbars1902380Nutrunner (high torque)
TABLE I : Battery Disassembly Task Breakdown
Fig. 3 : Left: Custom end-effector for the UR16e that holds a nut-runner and a Intel RealSense D435 in place. The nut-runner has been retrofitted with an Arduino-based interface board to control speed and measure current via Ethernet. Right: Block diagram of the Ethernet-based nutrunner control system.
Fig. 4 : Industrial workflow for setup (orange) and execution (green). Left: Relevant battery parts are labeled with text (1) and trained (2). All part locations and disassembly sequence are stored in JSON description (3) and right. Middle: Execution starts with a full scan, registration of parts (A), and uses agentic AI to execute the plan together with a human co-worker (B). Right: Sample JSON data structure.
IK MethodCandidatesSR (%) ↑\uparrowAvg. Plan Time (s) ↓\downarrow
KDL, 1 seed125.71.47
KDL, 50 random seeds≤\leq 5073.010.46
Analytical (ours)≤\leq 16899.30.31
TABLE II : IK solver comparison over 1 000 random goal poses with a 20 s planning budget.
 YoloWorldX [ 18 ]Yolov11L [ 25 ]
mAP0.50.97570.9708
mAP50-950.76250.7336
Precision0.92060.9359
Recall0.96730.9244
TABLE III : Object recognition performance of an LLM-ready open-world ready model (left) vs. class-based vision system (right).
Fig. 5 : Confusion matrix for YoloWorld on test set containing 18 Bolts, 27 BusBars, 85 InteriorScrews, 40 Nuts, 48 Orange Covers, 69 Screw, and 71 Screw Holes
 Taught-in Poses1-shot visionVisual Servo
Avg Time (min)24.0928.7036.29
Success rate97.06%57.35%82.84%
TABLE IV : Success rate for different removal strategies (n=204).
Fig. 6 : Metrics across three LLM back-ends for agentic execution of simple robot-motion commands “Move the robot 0.2m forward/backward” via MCP (orange) and an explicit SmolAgents tool (blue). Boxes: Q1/median/Q3; whiskers: 1.5 IQR; circles: outliers; dots: individual trials ( n=20n=20 ).
Fig. 7 : Metrics across three LLM back-ends for agentic execution of the complex screw-removal task (identify 5 of 68 remaining parts from the database, remove them, and update status) via MCP (orange) and an explicit SmolAgents tool (blue). Boxes: Q1/median/Q3; whiskers: 1.5 IQR; circles: outliers; dots: individual trials ( n=10n=10 per approach). Although performance looks similar, MCP has a 43.3% failure rate vs. 0% for the tool-based approach.
More in this topic

Robotic Manipulation

Robotic manipulation studies how robots perceive, plan, and control physical interactions to grasp, move, assemble, and use objects.

See all 1,615 papers

Explore similar work

CardsList
Aug 21, 2026

GOLEM: Modular Humanoid Autonomy Towards Electric Vehicle Battery Disassembly

Disassembling end-of-life electric vehicle (EV) battery packs is dull and dangerous work, performed almost entirely by humans. We present GOLEM (Generalized Open Library of Embodied Modules), an end-to-end, open-source system architecture for EV battery disassembly with the Unitree H1-2 humanoid robot in which walking, manipulation, dynamic stability, navigation, and spatial memory are independent modules with abstract interfaces, so that methods are easily developed, interchanged, and compared. GOLEM is deployed as a Docker-based ROS 2 abstraction in which MuJoCo and IsaacLab digital twins expose interfaces matching the physical robot. GOLEM's composability and per-module customization enable development and demonstration of humanoid EV battery disassembly, from simulation to reality. GOLEM provides fair comparison between humanoid modules, enabling evaluation as a capability ladder, in which one module is characterized at a time and added as a rung: LiDAR-inertial navigation places the robot within 13.0cm of a 6m goal; a learned standing controller recovers from external disturbances that sampling-based lower-body MPC does not; and grasping loosened fasteners from a real Hyundai Ioniq 5 pack degrades from 97% tethered to 87% free-standing to 37% under navigation-induced pose variance. Source code is available at the project page https://golem-humanoid.github.io

Max Conway, William Xie, Allen Devaraj +14

Department of Computer Science, University of Colorado Boulder, USA. · Department of Aerospace and Mechanical Engineering, University of Notre Dame, USA.

Modular RoboticsHumanoid Robot ControlRobotic ManipulationHumanoid Robot Locomotion
AskPDF
Jun 6, 2026

Vision-Guided Dual-Arm Humanoid Robotic Disassembly of End-of-Life 18650 Lithium-ion Battery Packs

The growing volume of retired lithium-ion battery packs from electric vehicles and portable electronics calls for automated disassembly that is safe, flexible, and selective down to the individual cell. Existing robotic systems, however, mostly assume known pack poses, external fixtures, or specialised tooling, leaving fixture-free cell-level disassembly under pose uncertainty largely unsolved. This paper presents a vision-guided dual-arm pipeline that disassembles a 21-cell 18650 pack from an arbitrary initial pose using only general-purpose parallel-jaw grippers, RGB-D sensing, and a pre-trained grasp detector. Pose uncertainty is absorbed by a learn-and-filter perception stack with discrete look-and-move wrist-camera corrections, while a mid-task support transfer between the two arms extends the effective workspace without any external clamp. The pipeline achieves an 8/10 end-to-end success rate, a cell-localisation root-mean-square error of 2.42.4,mm, and a mean cycle time of 6.0,minutes per pack, providing a practical, fixture-free building block for industrial battery recycling.

Yile Chen, Zhihao Liu, Xi Vincent Wang +1

KTH Royal Institute of Technology, Brinellvägen 8, 114 28, Stockholm, Sweden

Bimanual Robotic ManipulationRobotic ManipulationIndustrial RoboticsVision-Based Robot ControlLithium-Ion BatteriesHumanoid Robotics
AskPDF
Mar 10, 2026

SELF-VLA: A Skill Enhanced Agentic Vision-Language-Action Framework for Contact-Rich Disassembly

Disassembly automation has long been pursued to address the growing demand for efficient and proper recovery of valuable components from the end-of-life (EoL) electronic products. Existing approaches have demonstrated promising and regimented performance by decomposing the disassembly process into different subtasks. However, each subtask typically requires extensive data preparation, model training, and system management. Moreover, these approaches are often task- and component-specific, making them poorly suited to handle the variability and uncertainty of EoL products and limiting their generalization capabilities. All these factors restrict the practical deployment of current robotic disassembly systems and leave them highly reliant on human labor. With the recent development of foundation models in robotics, vision-language-action (VLA) models have shown impressive performance on standard robotic manipulation tasks, but their applicability to complex, contact-rich, and long-horizon industrial practices like disassembly, which requires sequential and precise manipulation, remains limited. To address this challenge, we propose SELF-VLA, an agentic VLA framework that integrates explicit disassembly skills. Experimental studies demonstrate that our framework significantly outperforms current state-of-the-art end-to-end VLA models on two contact-rich disassembly tasks. The video illustration can be found via https://zh.engr.tamu.edu/wp-content/uploads/sites/310/2026/03/IROS-VLA-Video.mp4.

Chang Liu, Sibo Tian, Xiao Liang +1

J. Mike Walker ’66 Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA. · Zachry Department of Civil and Environmental Engineering, Texas A&M University, College Station, TX 77843, USA.

Vision-Language-Action ModelsContact-Rich Robotic ManipulationRobotic ManipulationLong-Horizon Robotic Manipulation
AskPDF