“This is a proxy for contested logistics… The Baja Desert provides us with adverse terrain topography and weather; it also provides us a dynamic [operational tempo] so we can't pre-predict or plan anything,” said Brandon Bean, GDIT’s vice president for artificial intelligence and machine learning.
Project Celerity: GDIT and AWS field an energy-aware AI platform
General Dynamics Information Technology (GDIT) and Amazon Web Services (AWS) are partnering on Project Celerity, an “AI-enabled platform for managing energy” that the companies will exercise during the upcoming Baja 1000. The work is explicitly positioned as a stepping stone from a civilian motorsport environment to military use: GDIT says lessons learned in the thousand-mile race will inform future deployments “before it heads off to rough and disconnected battlefields.”
The platform ties energy management to logistics planning: microgrids and battery systems under study by the Army’s Advanced Research Lab are meant to power not only bases but the batteries for “a growing fleet of eclectic and robotic vehicles and weapons.” Project Celerity’s aim, as described in the source material, is to predict when and how vehicles such as drones or ground robots might need new batteries and to schedule maintenance decisions accordingly.
DOGMA: fusing sensor data and operating under degraded communications
The effort builds on GDIT’s Defense Operations Grid-Mesh Accelerator (DOGMA), introduced last August during the Pentagon’s T-REX drone warfare experiment. Bean said the company has developed three versions of DOGMA: one for fusing sensor data, one for running autonomy, and a layer called WorldView, which he described as a “cognitive layer that provides a common operating picture.”
DOGMA’s stated design is to operate where communications are limited or contested: the tool “fuses sensor data and sends it back to an operator under difficult conditions, such as enemy jamming, broken communications links, etc.,” according to the reporting. GDIT’s WorldView is the interface that consumes telemetry and patterns to give operators situational awareness despite those constraints.
Baja 1000 deployment: electrically powered bikes, telemetry, and pit predictions
Pilot Racing, a team in the Baja 1000, will use electrically powered bikes and carry the Project Celerity technologies onto the course. The race crosses California and Mexico and stretches roughly one thousand miles—GDIT calls it “the perfect” environment to stress-test predictive logistics and maintenance AI.
Bean said the bikes closely resemble electrically powered machines used in some special operations missions: they are quieter than conventional motorbikes and their large batteries can power sensors and communications gear. “All the telemetry that's coming from the rider and from the motorcycle” will be routed to AWS servers, Bean said, and predictive analytics will advise “when and where the rider needs to pit and where we need to replace the batteries.”
Rider-health telemetry, pattern-of-life, and expanding sensor inputs
GDIT and AWS are not limiting themselves to vehicle-state data. The companies have unveiled a rider-health tool intended for “no-communication environments where standard fitness trackers don’t work,” and demonstrated DOGMA WorldView and that health capability at SOF Week in Tampa, Florida.
Bean described a “round-loop workflow” where telemetry from devices is pulled into DOGMA WorldView to perform pattern-of-life analysis: “So we could tell, based on the telemetry data on the phone, whether they've [encountered] elevated terrain or whether they stopped for periods of time.” He added that the next step is to “actually tap into the microphone and the camera on the phone, so that we can identify if there's hostile control [of the] device.”
What this means for technologists, procurement officers, and field operators
- Technologists and security teams: Expect integration work across energy-management AI, autonomy runtimes, and data-fusion layers such as DOGMA. They will have to handle telemetry streaming to cloud servers (AWS, in this case) and consider how additional sensor inputs—microphone and camera—are integrated into pattern-of-life analytics.
- Procurement officers and program managers: The Baja 1000 exercise is being used as a proxy for “contested logistics.” Procurement teams will be watching how predictive analytics for pit timing and battery replacement perform in a highly dynamic environment that resembles denied or degraded battlefield conditions.
- Field operators and tactical units: The deployment emphasizes quieter electric mobility that can power communications and sensors. Operators will be testing whether predictive logistics and a cognitive WorldView can make reliable maintenance and pacing decisions where connectivity is limited.
The short-term objective is straightforward: use a demanding endurance race to stress test energy-aware AI, data fusion, and autonomy under real operational variability. The broader claim is equally explicit in the source material—these capabilities are being honed as a precursor to use “on rough and disconnected battlefields” and for noncombat missions ranging from patrols in remote areas to disaster or humanitarian response, as Shannon Judd, the director of global defense partners at AWS, wrote in an email.
Whether the Baja run becomes a proving ground that directly shapes fielded logistics AI will depend on the telemetry performance, the DOGMA WorldView’s ability to synthesize patterns in denied conditions, and how additional sensor inputs are managed when offline. For now, the exercise places a power-management problem—when a vehicle needs a battery and a pit stop—at the center of a larger experiment in distributed, contested operations.
Source: Defense One — Desert e-bike race ‘the perfect’ place to test military-vehicle AI
