The value of information and the emergence of communication in evolved e-puck robots
AuthorsFernando Montes-Gonzalez
AffiliationsInstituto de Investigaciones en Inteligencia Artificial, Universidad Veracruzana, Mexico
In simulated robot teams, communication evolves when sharing information helps find food, while innate reflexes can get in the way.
Key results
Controller size for each robot.
Mean champion fitness across three runs in the small arena.
Mean champion fitness across three runs in the small arena.
Champion fitness; exceeded solitary search by 0.165.
What the paper found
This study asks when robot communication is useful and whether innate action-selection rules help or hinder it. Two e-puck robots search for food they can detect only by passing over it, using controllers evolved for food-on-target time rather than for signaling. Each controller is a 16-unit Res-GRU with 1,768 parameters and an optional one-bit signal whose arrival direction the receiver can sense; the robots run in a PyBullet model validated against Webots. Without a compass, three independent runs converged on different strategies: a signal meaning “I’m on the food,” a beacon, or no signaling. Giving the robots a compass made the signal disappear, showing that communication was selected when it carried information unavailable through another channel. In the small arena, the Res-GRU condition reached mean fitness of 0.587, compared with 0.362 for EASA, which uses basal-ganglia-inspired action selection. In a larger arena, the referential signal scored 0.551 and beat solitary search by 0.165 fitness, while the beacon performed worse than searching alone. The larger task did not increase how often communication evolved; it made the usefulness of different signals clearer. EASA’s innate avoidance and exploration behaviors disrupted cooperation, and evolution consistently reduced their influence. Overall, the results suggest that communication depends on the information a signal adds, while an architecture designed around isolated-agent reflexes can constrain performance in a social task.
Original abstract
We study when communication emerges between autonomous agents and what role an innate action selection architecture plays. Two e-puck robots, reconstructed in PyBullet from their Webots model and validated against it, must find a feeding zone that can only be detected when a robot passes over it. Each robot can emit a one-bit signal whose direction is detected by the receiver. Control is provided by a GRU recurrent network with a residual path, either operating alone or as a channel within an Embedded Action Selection Architecture (EASA). Fitness measures only the time the robots spend over the food. When the robots receive a compass pointing to the food, use of the signal disappears completely. Without a compass, in a small arena, three independent runs achieved similar performance using different strategies: a referential signal, a permanent beacon, or no use of the signal. As the arena expands, the strategy based on a referential signal generalizes best. In contrast, evolving directly in the large arena does not increase the frequency of communication, but it does separate the strategies, and the referential signal clearly outperforms solitary searching. Within EASA, innate reflexes constrain the network and reduce performance. As its priorities evolve, evolution consistently weakens those reflexes. The results suggest that communication emerges in proportion to the value of the information it transmits, and that an innate scaffolding designed for isolated agents can be an impediment in a social task.
Read the original paperMore in Robotics
Browse all 53 papers →Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration
Jusuk Lee, Sungha Kim, Yeonsoo Park, Jonguk Cheon, Yoonkyo Jung, Yongjun You, H. Jin Kim, Jia-Bin Huang, Furong Huang, Youngseok Jang, Seungjae Lee
A single human video is transformed into scene-graph guidance that helps a robot hand learn dexterous skills and adapt to situations it never saw.
Physical Time as a Missing Dimension of Robot Policy Memory Dynamics
Yushen Liang, Yue Peng, Baosheng Jin, Tianluo Zhang, Xinyu Zhang, Shuyi Zhou, Zhuoran Chen, Xinqi Liu, Shenji Wan
Temporal Flow Policy gives robot memories a clock, helping actions reflect how much real time has passed.
JAMB: Joint Action-Motion Diffusion for Bimanual Manipulation
Chuyang Xiao, Peilin Meng, David Held
JAMB helps two robot arms coordinate by jointly imagining their future movements and the changing 3D scene before acting.