Baja 1000 E-Bike Competition Serves as Ideal Testing Ground for Military AI Systems

A participating team in the upcoming Baja 1000 endurance bike event will field an AI system that determines optimal pit‑stop timing well before it becomes critical.

The 1,000‑mile route spanning California and Mexico provides an ideal proving ground for GDIT’s logistics and maintenance AI, which will later be adapted for austere battlefield conditions with limited infrastructure.

GDIT is collaborating with AWS on Project Celerity, an AI‑driven platform designed to manage energy resources. The Army’s Advanced Research Laboratory has prioritized the deployment of compact tactical microgrids—integrated generation and storage units that operate in off‑grid environments. Beyond powering personnel, these microgrids will support batteries for an expanding array of autonomous vehicles and robotic platforms, making accurate forecasts of battery demand a critical component.

Shannon Judd, AWS Director of Global Defense Partnerships, explained in an email that Project Celerity has broad military utility, supporting remote patrols, special‑operations decision‑making in communications‑deprived settings, and disaster‑relief or humanitarian operations, as well as broader infrastructure management.

Brandon Bean, GDIT’s Vice President of AI and Machine Learning, noted, “This initiative serves as a testbed for contested logistics. The Baja desert’s rugged terrain, variable weather, and unpredictable operational tempo prevent pre‑planned solutions.” GDIT has not disclosed which specific team will employ its technology.

This effort builds on GDIT’s Defense Operations Grid‑Mesh Accelerator (DOGMA), a sensor‑fusion system that delivers real‑time data to operators even under challenging conditions such as jamming or disrupted communications.

Since its debut in August during the Pentagon’s T‑REX drone experiment, GDIT has refined DOGMA into three distinct versions: a data‑fusion module, an autonomy‑execution module, and WorldView—a cognitive layer that aggregates a unified operational picture.

The race team will compete on electric‑powered bicycles reminiscent of those employed by certain special‑operations units, which are quieter than conventional motorbikes and can additionally supply power to onboard sensors and communications equipment.

Bean explained that all rider and motorcycle telemetry will stream to AWS servers, enabling predictive analytics to identify optimal pit‑stop locations and the precise moments when battery replacements are required.

Future extensions may incorporate a rider‑health monitoring tool tailored for environments lacking connectivity, where conventional fitness trackers are ineffective. The company demonstrated this capability alongside DOGMA WorldView at the recent SOF Week gathering in Tampa, Florida.

Bean described the creation of a closed‑loop workflow that harvests telemetry from the devices, feeds it into DOGMA WorldView, and enables pattern‑of‑life analysis. This allows inference of terrain changes or prolonged stops from phone telemetry, and the next phase involves leveraging the phone’s microphone and camera to detect hostile control of the device.

Source link