Dark Mode


Creating Formations with Multiple Agents



Formation shapes

Figure 6
: We extend our method to agents coordinating and forming various shapes. Various shapes are created using a set of ”expected positions” around one or two landmark positions. The agents use these expected positions as goals.

Comparison over 100 episodes

3agent results

(a) With 3 agents.
5 agent results

(b) With 5 agents
10 agent results

(c) With 10 agents
Figure 7: Circle Formation: The violin plots show the distribution of fairness (𝓕) and the total distance traveled (D) over 100 test episodes for three trained model variants: 1) Optimal distance cost goal assignments (OA); 2) Fair goal assignments (FA); and 3) Fair goal assignments and a fairness reward (FA+FR). A white circle and tick denote the medians; a plain tick represents the means, and the vertical black lines indicate the 90-10 percentile range.

Performance in Congested Environments

congestion envs

Figure 8
: Congestion in the environment: The figure on the left shows an environment with 3 agents along with 3 obstacles and 2 walls. The figure on the right shows the environment with 7 agents and 3 obstacles. The environment is crowded with the increased number of agents, which decreases free space for navigating in straight lines.


Comparison over 100 randomly initialized congested environments

3 agent results

(a) With 3 agents.
5 agent results

(b) With 5 agents
10 agent results

(c) With 10 agents
Figure 9: Congestion: The violin plots show the distribution of fairness (𝓕), the total distance traveled (D) and success rates (S%) over 100 test episodes for three trained model variants: 1) Random goal assignments (RA); 2) Optimal distance cost goal assignments (OA); 3) Fair goal assignments (FA); and 4) Fair goal assignments and a fairness reward (FA+FR). A white circle and tick denote the medians; a plain tick represents the means, and the vertical black lines indicate the 90-10 percentile range. We also show the tradeoffs between fairness and efficiency exhibited by the different models in the rightmost subplot.

Conclusions


Future Work



 [Slides]

Click on the link if PDF is not rendered.

Citation


If you find our work or code useful in your research, please consider citing the following paper:

We have extended this approach to work with more congested environments and different formation shapes.

Related Links

Other related papers and works

Acknowledgments


The authors would like to thank the MIT SuperCloud and the Lincoln Laboratory Supercomputing Center for providing high-performance computing resources that have contributed to the research results reported within this paper. This work was supported in part by NASA under grant #80NSSC23M0220 and the University Leadership Initiative (grants #80NSSC21M0071 and #80NSSC20M0163), but this article solely reflects the opinions and conclusions of its authors and not any NASA entity. J.J. Aloor was also supported in part by a Mathworks Fellowship.

This website template was edited heavily following the base from Michaël Gharbi and Matthew Tannick.