From Stop Requests to Safe Handoffs: Verifying the Action Boundary in Gwendolen Robot Teams

When a robot team transfers a task, a controller's acceptance of a cancellation request can precede the end of the action. This study examines what evidence a team needs before another robot starts. It combines bounded model checking of Gwendolen agent programs with 7,350 simulated rescue missions paired by seed. Under truthful action reports and authentic grants, the handoff contract satisfied the safety properties in every completed two-robot search. No three-robot coalition safety search completed. Where a terminal result already implied the safe state, standard correlated cancellation, terminal acknowledgement and retirement produced the same simulated outcomes. Waiting for separate safe-state evidence delayed handoffs when the two events were separated. Loss of terminal results reduced completion for both approaches. Experiments with intention scheduling also identified a race missed by passing searches. A separate tool-call study illustrates how a durable service effect can survive cancellation. The robot results describe bounded verification and simulation.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24374367
Primary Topic
Formal Methods in Verification
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

From Stop Requests to Safe Handoffs: Verifying the Action Boundary in Gwendolen Robot Teams

S. M. Abdulla Al Mamun
Zenodo (CERN European Organization for Nuclear Research)
Formal Methods in Verification
preprint

From Stop Requests to Safe Handoffs: Verifying the Action Boundary in Gwendolen Robot Teams

S. M. Abdulla Al Mamun
preprint en

Abstract

When a robot team transfers a task, a controller's acceptance of a cancellation request can precede the end of the action. This study examines what evidence a team needs before another robot starts. It combines bounded model checking of Gwendolen agent programs with 7,350 simulated rescue missions paired by seed. Under truthful action reports and authentic grants, the handoff contract satisfied the safety properties in every completed two-robot search. No three-robot coalition safety search completed. Where a terminal result already implied the safe state, standard correlated cancellation, terminal acknowledgement and retirement produced the same simulated outcomes. Waiting for separate safe-state evidence delayed handoffs when the two events were separated. Loss of terminal results reduced completion for both approaches. Experiments with intention scheduling also identified a race missed by passing searches. A separate tool-call study illustrates how a durable service effect can survive cancellation. The robot results describe bounded verification and simulation.

Zenodo (CERN European Organization for Nuclear Research)
University of Manchester (GB)
Formal Methods in Verification
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon