Robot memory improves by tracking physical time and past events
Physical Time as a Missing Dimension of Robot Policy Memory Dynamics
Robotics
Summary
Robots usually remember things based on the order of what they see, assuming each moment passes equally fast. But the authors show that the real passage of time matters for how robots should act because the same events could take different amounts of time. They made a new way for robots to keep memories that change depending on how much time has passed, which helps robots better decide what to do next. Tests showed this method helps robots perform tasks that need careful timing or remembering even after not seeing something for a while.
What this means in practice
- •For robot developers: Create robot controllers that adjust actions based on how much real time elapsed to improve timing and memory in tasks.
- •For industrial automation engineers: Improve robots handling complex physical tasks by incorporating measured time intervals into memory and decision-making.
Authors
Yushen Liang, Yue Peng, Baosheng Jin, Tianluo Zhang, Xinyu Zhang, Shuyi Zhou, Zhuoran Chen, Xinqi Liu, Shenji Wan
Abstract
Existing robot policy memories typically evolve with observation sequences, implicitly treating each update as the same amount of physical time. Yet identical observation histories can span different durations and require different actions, exposing physical time as a missing dimension of policy memory. We introduce Temporal Flow Policy (TFP), which learns memory dynamics to retain past events and evolve the memory state over measured time intervals. Observations and prior memory determine the content and time scales of each transition; elapsed time determines how far the state advances. TFP conditions diffusion and flow-matching policies through this memory state and is trained with their native imitation objectives. We evaluate TFP with Diffusion Policy and $π_{0.5}$ through controlled comparisons of memory designs. These experiments show that successful timing depends on both measured intervals and retained history. TFP achieves higher average success than the recurrent and state-space memories we equip with the same intervals and than a variant that feeds the interval to a learned gate. Memory readouts recover the time since an event and motion information after occlusion. Changing the supplied intervals systematically delays or advances actions, linking time inputs to behavior. Real-robot tasks requiring timing or memory, together with LIBERO evaluations, demonstrate benefits in physical and general manipulation. These results suggest viewing robot memory not as stored observation history, but as a state that retains past events and evolves with physical time.