MeshRTC-Agents: Software Agents for Robotic Transport Platforms Coordinated by Short Meshtastic Commands — Synthetic Digital Twin, Neural Modules and Google Colab Notebook (v1.1.0)

Description During prolonged power outages, ground robots that deliver medicines or drinking water may lose contact with their dispatcher mid-mission. This package provides software agents for such robotic transport platforms, coordinated through short commands over a Meshtastic LoRa mesh. It contains: A compact protocol of fifteen commands. They cover a task contract, status, decision, blockage, receipt, site closure, request, auction bid, lease, recall, relay deployment, risk dial, reporting policy, hold and acknowledgement. Each command has a text and a binary (CRC-16) encoding and passes a deterministic validator. The package also models LoRa airtime and the gateway's 1% duty-cycle budget. A synthetic digital twin of a riverside town. It uses an Okumura–Hata radio model, cellular coverage that fades with outage age, and solar repeaters. Agents. A crisis-coordination orchestrator, a comms steward, robot agents, recipient nodes and a relay UGV. The robot agents use one of three strategies: stop-on-loss, fixed-route, or a Supervisory Feasibility Controller with chance-constrained energy checks. Fleet logic with store-carry-forward exchange and in-field auctions. Ten transport scenarios. Four neural modules written in NumPy. An energy deep ensemble. A behaviour-cloned policy that runs under a runtime-assurance shield. A mesh-coverage predictor. An operator intent parser that turns Ukrainian or English text into validated commands. In 384 paired missions, the controller stranded no robot, against 72% (stop-on-loss) and 34% (fixed-route). Without the shield, a faulty neural network lost 21 of 45 robots; with it, none. The executed Google Colab notebook contains 39 figures and schematic diagrams. All data are synthetic; the town is inspired by, but is not a map of, Verkhnodniprovsk, Ukraine. Version 1.1.0 is the English edition of version 1.0.0, whose notebook text was in Ukrainian; code and results are unchanged. Keywords Meshtastic; LoRa; ground delivery robot; multi-agent systems; supervisory control; delay-tolerant networking; chance constraints; digital twin; neural networks; runtime assurance; emergency logistics; power outage

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-21
DOI
https://doi.org/10.5281/zenodo.24414186
Primary Topic
Robotics and Automated Systems
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

MeshRTC-Agents: Software Agents for Robotic Transport Platforms Coordinated by Short Meshtastic Commands — Synthetic Digital Twin, Neural Modules and Google Colab Notebook (v1.1.0)

Dmytro Prokopovych-Tkachenko
Zenodo (CERN European Organization for Nuclear Research)
Robotics and Automated Systems
article

MeshRTC-Agents: Software Agents for Robotic Transport Platforms Coordinated by Short Meshtastic Commands — Synthetic Digital Twin, Neural Modules and Google Colab Notebook (v1.1.0)

Dmytro Prokopovych-Tkachenko
article en

Abstract

Description During prolonged power outages, ground robots that deliver medicines or drinking water may lose contact with their dispatcher mid-mission. This package provides software agents for such robotic transport platforms, coordinated through short commands over a Meshtastic LoRa mesh. It contains: A compact protocol of fifteen commands. They cover a task contract, status, decision, blockage, receipt, site closure, request, auction bid, lease, recall, relay deployment, risk dial, reporting policy, hold and acknowledgement. Each command has a text and a binary (CRC-16) encoding and passes a deterministic validator. The package also models LoRa airtime and the gateway's 1% duty-cycle budget. A synthetic digital twin of a riverside town. It uses an Okumura–Hata radio model, cellular coverage that fades with outage age, and solar repeaters. Agents. A crisis-coordination orchestrator, a comms steward, robot agents, recipient nodes and a relay UGV. The robot agents use one of three strategies: stop-on-loss, fixed-route, or a Supervisory Feasibility Controller with chance-constrained energy checks. Fleet logic with store-carry-forward exchange and in-field auctions. Ten transport scenarios. Four neural modules written in NumPy. An energy deep ensemble. A behaviour-cloned policy that runs under a runtime-assurance shield. A mesh-coverage predictor. An operator intent parser that turns Ukrainian or English text into validated commands. In 384 paired missions, the controller stranded no robot, against 72% (stop-on-loss) and 34% (fixed-route). Without the shield, a faulty neural network lost 21 of 45 robots; with it, none. The executed Google Colab notebook contains 39 figures and schematic diagrams. All data are synthetic; the town is inspired by, but is not a map of, Verkhnodniprovsk, Ukraine. Version 1.1.0 is the English edition of version 1.0.0, whose notebook text was in Ukrainian; code and results are unchanged. Keywords Meshtastic; LoRa; ground delivery robot; multi-agent systems; supervisory control; delay-tolerant networking; chance constraints; digital twin; neural networks; runtime assurance; emergency logistics; power outage

Zenodo (CERN European Organization for Nuclear Research)
University of Customs and Finance (UA)
Robotics and Automated Systems
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