Synthesizing and revising plans for autonomous robot adaptation

AI Planning is concerned with producing plans that are guaranteed to achieve a robot’s goals, assuming  the pre-specified assumptions about the environment in which it operates hold.  However, no matter how detailed these assumptions are or how complex the resulting plan is, unexpected events at runtime can cause the robot to fail in achieving its goals.

For robots to safely adapt when such events occur, they must be able to reassess 1) the assumptions upon which they depend, 2) the achievability of its goals under the new environment conditions and 3) the quality of the plan that will achieve the goals or their variants.
 
Symbolic learning and AI planning have studied these problems in isolation and can be enable robots to adapt sequentially and iteratively; iteratively revise the assumptions until goals are realizable and then synthesize a plan. However, such a procedure can be costly and lack convergence guarantees (e.g.,  convergence may require more iterations than reasonable for a robot waiting to decide how to proceed in the face of an unexpected event).
 
This PhD will focus on the theory and implementation of a new hybrid symbolic learning and planning approach for assured runtime adaptation in the face of unexpected events. Specifically, it will look at developing a meta symbolic learning approach that uses knowledge about candidate specifications (e.g., weakness and realisability), alternative adversary models for controller synthesis and previous adaptations to guide its search for suitable adaptations.

Dalal Alrajeh, Antoine Cailliau, Axel van Lamsweerde:
Adapting requirements models to varying environments. ICSE 2020: 50-61
Davide G. Cavezza, Dalal Alrajeh, András György:
A Weakness Measure for GR(1) Formulae. Formal Aspects Comput. 33(1): 27-63 (2021) Pawel Gomoluch, Dalal Alrajeh, Alessandra Russo, Antonio Bucchiarone:
Learning Neural Search Policies for Classical Planning. ICAPS 2020: 522-530
Daniel Alfredo Ciolek, Víctor A. Braberman, Nicolás D’Ippolito, Sebastian Sardiña, Sebastián Uchitel:
Compositional Supervisory Control via Reactive Synthesis and Automated Planning. IEEE Trans. Autom. Control. 65(8): 3502-3516 (2020)
Leandro Nahabedian, Víctor A. Braberman, Nicolás D’Ippolito, Shinichi Honiden, Jeff Kramer, Kenji Tei, Sebastián Uchitel:
Dynamic Update of Discrete Event Controllers. IEEE Trans. Software Eng. 46(11): 1220- 1240 (2020)
Nicolás D’Ippolito, Víctor A. Braberman, Daniel Sykes, Sebastián Uchitel:
Robust degradation and enhancement of robot mission behaviour in unpredictable environments. In Proceedings of the 1st International Workshop on Control Theory for Software Engineering (CTSE 2015). Association for Computing Machinery, New York, NY, USA, 26–33. DOI:https://doi.org/10.1145/2804337.2804342

Project ID

STAI-CDT-2021-IC-20

Supervisor

Dalal Alrajehwww.doc.ic.ac.uk/~da04

Category

AI Planning, Logic, Verification