Prediction-Error-Tightened MPC for Socially Compliant Crowd Navigation
Organizations: Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioingegneria, Milan, Italy · Robotics Laboratory, Innovation Hub & Intellectual Property, Leonardo S.p.A., Genoa, Italy
Abstract
Autonomous navigation of mobile robots in crowded environments, while preserving pedestrians' psychological comfort, requires distinguishing pedestrians from other obstacles, predicting their future motion, and accounting for uncertainty in the predicted trajectories of both pedestrians and the robot. This work extends a previous MPC framework by integrating obstacle and pedestrian detection, Koopman-based motion prediction, and online prediction-error estimation directly within the collision-avoidance mechanism. To reduce collision-avoidance complexity in crowded scenarios, nearby obstacles are aggregated, thereby reducing the number of collision-avoidance constraints. Simulation studies evaluate the prediction-error margins and the computational benefits of semantics-aware aggregation. In a paired 300-seed closed-loop comparison with a moving generic obstacle, Markov residual tightening reduces the fraction of runs with physical contact from to , at a mean travel-time cost of s. In a separate crowd-navigation semantic ablation, the offline robot-footprint/personal-space overlap occupies of observed time with pedestrian-aware handling versus with generic-obstacle handling.
Figures & tables
| Mean VR [%] | Variance [% 2 ] | Valid runs | ||||
|---|---|---|---|---|---|---|
| Configuration | Ped. | Generic | Ped. | Generic | Ped. | Generic |
| 52.33 | 52.21 | 110.41 | 135.95 | 293 | 284 | |
| 1.04 | 0.90 | 0.73 | 0.70 | 299 | 287 | |
| 2.39 | 2.37 | 2.63 | 3.04 | 297 | 289 | |
| 4.99 | 5.45 | 7.06 | 11.00 | 299 | 282 | |
| 7.38 | 8.18 | 11.51 | 17.45 | 299 | 286 | |
| Metric | Markov | |
|---|---|---|
| Contact runs | ||
| Nav2 goal success | ||
| Time to goal [s], mean | 23.00 | 22.32 |
| Path length [m], mean | 11.296 | 11.006 |
| Minimum planar clearance [m], mean | 1.238 | 0.989 |
| Failed QPs / post- solves |
| Case | Pairs | Constraints | Mean [ms] | p99 [ms] | |
|---|---|---|---|---|---|
| A | 4 | 524 | 105/143 | 12.35/12.95 | 56.0/67.0 |
| A | 8 | 539 | 127/199 | 25.85/28.71 | 106.5/136.1 |
| A | 16 | 478 | 162/290 | 54.57/71.38 | 255.7/358.9 |
| B | 4 | 418 | 122/158 | 9.18/9.99 | 32.3/33.1 |
| B | 8 | 637 | 136/217 | 11.35/13.11 | 40.6/51.1 |
| B | 16 | 987 | 160/294 | 17.61/20.69 | 66.0/100.6 |
| Pedestrians | Valid runs | Pedestrian [%] | Generic [%] |
|---|---|---|---|
| All |
Social Robot Navigation
Social robot navigation studies how robots move safely and comfortably around people while respecting human behavior, preferences, and shared-space norms.
See all 87 papers