Kairos predicts future people presence and movement in 3D spaces
Kairos: Grounded Forecasting of Presence and Directional Flow in 4D Scene Graphs
RoboticsArtificial Intelligence
Summary
Robots working in busy places need to guess where people might be and how they might move in the future. Existing methods either simplify movement to a point or keep movement patterns fixed over time. The authors created Kairos, a system that remembers how people move through 3D maps over time, predicting not just where people will be but also the directions they will travel at any future time. They tested it using real data from places like a campus, mall, and train station, showing it can make useful predictions from limited robot observations. This helps robots plan paths that better avoid people or interact with them safely.
What this means in practice
- •For robot developers: Enable robots to predict where people will be and their movement direction to improve navigation and safety in busy environments.
- •For urban planners: Use directional pedestrian forecasts from limited observation data to analyze and design public spaces for better flow and crowd management.
Authors
Iacopo Catalano, Julio A. Placed, Javier Civera, Jorge Peña Queralta
Abstract
Long-term autonomy in human-populated environments requires anticipating whether and how people will move at times a robot has not yet observed. Existing representations of pedestrian motion face a tradeoff: they either forecast future activity, reducing each location to a scalar rate, or model the full directional distribution, holding it fixed in time. We present Kairos, a predictive directional-flow memory that extends a hierarchical 3D scene graph (3DSG) to a 4D scene graph (4DSG). Every observed voxel of the reconstructed geometry stores a directional mixture and a presence rate, and spectral predictors forecast, for any future query time, both the probability that people are present and the full directional distribution of their motion. Pairwise flow dependence between adjacent voxels supports conditional queries, and per-voxel predictive variances yield calibrated credible intervals that tighten as observations accumulate. We evaluate Kairos on three real pedestrian environments: a robot-collected campus dataset, a shopping mall, and a station concourse recorded continuously for eleven months. Its learned state remains consistent under loop-closure corrections, and its forecasts are competitive with dedicated occupancy and flow models trained on the full detection stream, although Kairos learns from only the small fraction available to a patrolling robot. Finally, we validate the representation on a downstream encounter-probability planning task, where plans computed over the Kairos forecasts encounter more people than plans computed over any time-invariant map at an equal success rate. We provide the code at https://github.com/IacopomC/kairos.