Autonomous Whole-body Loco-manipulation Using a Snake Robot

Adarsh Salagame; Henry Noyes; Haroon Hublikar; Alireza Ramezani

cs.RONov 6, 2026

Autonomous Whole-body Loco-manipulation Using a Snake Robot

Authors: Adarsh Salagame, Henry Noyes, Haroon Hublikar, Alireza Ramezani

Organizations: Department of Electrical and Computer Engineering, Northeastern University, Boston MA

Abstract

Snake robots can leverage their hyper-redundant morphology for both locomotion and manipulation through distributed body contact but whole-body loco-manipulation is particularly challenging since the entire body serves simultaneously as propulsive mechanism and manipulation tool Prior work on wheeled legged and snake robot platforms has decoupled these functions through dedicated appendages body segmentation or grippers We present closed-loop loco-manipulation with our snake robot COBRA without this distinction such that any part of the body may make contact We develop a contact-implicit dynamic model unifying robot-ground robot-object and object-ground interactions within a Moreau-Jean time-stepping framework and formulate a trajectory optimization over the full-order dynamics guided by a geometric planner and nominal proportional controller to autonomously push an object to a goal location during sidewinding We validate the planner and controller on hardware across varied configurations and present simulation results for a simplified trajectory optimization that resolves robot-object and object-ground contacts

Figures & tables

Fig. 1: Shows COBRA demonstrating closed-loop loco-manipulation.

Fig. 2: Full-order and reduced-order representations of the snake robot during loco-manipulation. Left: full-order model showing floating-base pose $\bm{q}_{b}$ , joint coordinates $\bm{q}_{j}$ , robot–ground contact impulses $\bm{\Lambda}_{rg}$ with friction cone $K_{\mu}$ , and bounding box projection $\bm{p}_{bb}$ . Right: reduced-order planning space with robot start pose (R), box start pose (B), and goal pose (G). The robot reference trajectory $\bm{q}_{bb}^{\mathrm{des}}$ consists of an approach phase and a loco-manipulation phase. The desired box trajectory $\bm{q}_{\mathrm{box}}^{\mathrm{des}}$ defines the box path to the goal. The robot–box contact impulse $\bm{\Lambda}_{rb}$ acts at the contact point subject to friction cone $K_{\mu_{r}}$ .

Fig. 3: Planar free body diagram of the box. The robot applies force $\bm{F}$ at location $c_{x}$ on the active contact side. Ground contacts at $\pm l/2$ exert friction forces $\bm{\lambda}_{1}$ , $\bm{\lambda}_{2}$ .

Fig. 4: Reference trajectory generation in approach mode (left) and loco-manipulation mode (right). Approach uses a second-order Bézier with a single control point $\bm{c}_{1}$ along the heading $\bm{e}_{\theta_{bb}}$ . Loco-manipulation uses a third-order Bézier with control points $\bm{c}_{1}$ , $\bm{c}_{2}$ along the heading and robot–box directions. Both are placed at distance $\eta L$ from their respective anchor points.

Fig. 5: Hardware loco-manipulation results. Top left: composite snapshots of a representative trial showing robot start pose (R), box start pose (B), and goal region (G). Top right: corresponding trajectory plot with reduced-order bounding box, robot center-of-mass trajectory, box trajectory, and planned Bézier path. Color gradient indicates time progression. Bottom: trajectory plots from four additional trials with varied start and goal configurations.

Fig. 6: Composite snapshots from four hardware loco-manipulation trials corresponding to the trajectory plots in Fig. 5 (bottom). The red cross marks the goal position in each trial. Faded overlays show intermediate robot and box poses.

Fig. 7: Waypoint tracking experiment and contact-implicit model validation. Top left: composite snapshots showing approach, manipulation through waypoints WP1 and WP2, and delivery to the goal region. Top right: trajectory plot with robot trajectory, box trajectory, and reduced-order bounding box. Bottom left: robot–box contact force distribution against friction cone $K_{\mu_{r}}$ . Bottom center: box–ground friction forces $\bm{\lambda}_{1}$ , $\bm{\lambda}_{2}$ against friction polyhedron $K_{\mu_{g}}^{i}$ . Bottom right: box model validation via single shooting—simulated box trajectory overlaid on experimental trajectory with instantaneous pushing forces and ground friction forces at each time step.

Fig. 8: Reduced-order control signals during the waypoint tracking trial of Fig. 7 . Robot heading $\theta_{bb}$ , heading error $\theta_{bb}-\theta_{bb}^{\mathrm{des}}$ , and steering bias $\Delta$ . Vertical dashed lines mark box acquisition and waypoint transitions (WP1, WP2).

Contact-Implicit Trajectory Optimization

See all 9 papers

PAKE: Learning Whole-Body Loco-Manipulation with Partial Kinematic Embeddings
Jul 13, 2026Zhengmao He, Moonkyu Jung, Hyeongjun Kim +4Normalizing Flows Whole-Body Robotic Manipulation
The Hong Kong University of Science and Technology (Guangzhou). · Korea Advanced Institute of Science and Technology. · ETH Zurich. +1
Sumo: Dynamic and Generalizable Whole-Body Loco-Manipulation
Apr 9, 2026John Z. Zhang, Maks Sorokin, Jan Brüdigam +14Robotic Loco-Manipulation Generalization in Robotic Manipulation
MIT · RAI Institute · Cornell University
S-Rolling: Steerable Rolling Gaits for Twistable Snake Robots in Low-Clearance Environments
Oct 21, 2026Qingling Fan, Zhongqiang RenRobot Locomotion Robot Motion Generation
Global College, Shanghai Jiao Tong University, China

Jul 13, 2026J/K move · Enter open · S save

PAKE: Learning Whole-Body Loco-Manipulation with Partial Kinematic Embeddings

Zhengmao He, Moonkyu Jung, Hyeongjun Kim, Jiseong Lee+3

The Hong Kong University of Science and Technology (Guangzhou). · Korea Advanced Institute of Science and Technology. · ETH Zurich.+1

Normalizing Flows Whole-Body Robotic Manipulation Whole-Body Control Robot Motion Generation Robot Imitation Learning Robotic Loco-Manipulation

Open paper AskPDF arXiv

Loco-manipulation has recently shown promising capabilities; however, achieving high-precision control, managing the high-dimensional action space induced by many degrees of freedom (DoFs), and fully exploiting the inherent redundancy of whole-body systems remain challenging. In this paper, we propose a novel whole-body control framework that effectively addresses these challenges by decomposing the complex loco-manipulation problem into partial reference motion generation and low-level imitation control. We introduce a new Kinematic Normalizing Flow (KNF) model, trained on a large-scale kinematic dataset, that generates diverse partial kinematic references. A high-level controller is then trained to navigate the KNF's latent space to exploit redundant solutions, while a low-level controller learns dynamically consistent motion execution. We validate our approach on the quadrupedal robot equipped with a six-DoF robotic arm. In simulation, PAKE outperforms the evaluated baselines in tracking accuracy and feasible workspace coverage on the 4,000-target benchmark. For hardware deployment, we evaluate the system over 24 episodes across 8 different mobile loco-manipulation tasks. The system achieves end-effector pose-tracking errors of 4.5 cm and 0.14 rad, while maintaining accurate locomotion tracking with linear and angular velocity errors of 0.1 m/s and 0.01 rad/s, respectively. Our method represents a practical and powerful solution for accurate and generalized whole-body loco-manipulation in high-DoF robotic systems, with promising potential for diverse downstream robotic tasks.

Abstract

Contact-Implicit Trajectory Optimization

Explore similar work

PAKE: Learning Whole-Body Loco-Manipulation with Partial Kinematic Embeddings

Sumo: Dynamic and Generalizable Whole-Body Loco-Manipulation

S-Rolling: Steerable Rolling Gaits for Twistable Snake Robots in Low-Clearance Environments