Effects of Locomotion and Viewpoint Control on Immersive Robot Telepresence for Navigation and Inspection

TL;DR In a 17-person VR study, people teleoperated a real quadruped through a foam maze to test joystick vs omnidirectional-treadmill driving and joystick vs head-tracked camera control. Head-tracked camera control mattered most. Paired with joystick driving it finished fastest (178 s vs 258 s with joysticks for both) and had the fewest wall contacts (2.8 vs 6.1 per trial).
What they do
Inspecting a site through a mobile robot means steering and looking at the same time, and most VR telepresence interfaces put both on handheld controllers. The authors built VR telepresence for a Unitree AlienGo quadruped and ran a within-subject 2x2 study with 17 participants. They crossed locomotion (joystick, or walking on an Omnideck omnidirectional treadmill) with camera control (joystick, or head tracking). Participants drove the real robot through a 4.2 m x 9.0 m foam-wall maze and took screenshots of circle targets on the walls.
How it works
- Robot. A fixed stereo camera sits at the robot’s head and a second one on a servo pan-tilt mount above the torso. Stereo video streams to a Meta Quest Pro at 15 Hz.
- Locomotion. Joystick axes, or the user’s tracked translation and yaw on the treadmill, become bounded velocity commands (0.5 m/s, 1.5 rad/s), so every condition uses the same robot walking controller.
- Viewpoint. Head orientation or the right joystick sets pan and tilt. When both inputs are joysticks, a button toggles between driving and camera control.
- Study. Each participant ran 12 trials on 3 maze variants (204 trials in total). Wall contacts were logged through a Vicon-aligned Unity replica, and the analysis used mixed-effects models with Bonferroni-corrected comparisons.
Results
- Completion time: joystick driving + head camera 178 s, treadmill + head 214 s, joystick + joystick 258 s, treadmill + joystick camera 282 s.
- Wall contacts: 2.8 per trial for joystick + head vs 5.0-6.1 in the other conditions. Contact time was 11.6 s vs 19.5 s for joystick + joystick.
- With head-tracked camera control, operators switched views 1.6-1.9 times per trial instead of 6.4-7.5 and spent 83-85% of the time in the pan-tilt view instead of 63-65%.
- Workload (NASA-TLX) was lowest for joystick + head (3.76) and highest for treadmill + joystick camera (4.61). Target capture was 98.8-99.8% and sickness stayed low in every condition.
Limitations
The study had only 17 participants, and target capture was near ceiling, so it could not separate the interfaces. The task was a supervised, low-speed lab maze, and glass-to-glass video latency was not measured. The conditions mix posture, body coupling and controller allocation. The treadmill adds physical effort and limits who can use it. The camera-strategy analyses are exploratory and do not show causal mediation.
Figures



Abstract
In many robot telepresence interfaces, movement and viewing are managed through handheld inputs. This can divide attention between steering, looking, and inspection. We compared joystick versus omnidirectional treadmill locomotion and joystick-based versus head-guided camera control in a within-subject 2x2 study in which 17 participants used a virtual reality headset to guide a quadruped robot through a physical maze while locating marked wall symbols and capturing them as visual evidence. Mixed-effects analyses showed that head-guided camera control was associated with faster completion, lower wall-contact cost, lower workload, and more favorable presence, usability, technology acceptance, and flow than joystick-based camera control. The strongest overall objective performance was observed in the joystick locomotion plus head-guided camera condition, whereas the omnidirectional treadmill plus head-guided camera condition emerged as the strongest experiential compromise. Exploratory camera analyses showed that head-guided camera control reduced switching, increased reliance on the pan-tilt view, and supported a more stable viewing strategy across the maze. These strategy differences were consistent with the performance advantage, especially for wall-contact duration. Targets captured remained near ceiling and simulator sickness remained low across all conditions. Overall, the results indicate that viewpoint control was the stronger differentiator in this combined navigation and inspection task, while locomotion still shaped the objective and experiential tradeoff.
Summaries are written automatically from each paper's text. Figures and videos come from the paper and its project page; all credit goes to the authors.
