AI Edge Computing Drone Applications Course — 2 Week MTT

2-Week AI Edge Computing Drone Applications Course

This 2-week Mobile Training Team course is our most advanced offering. Students build sUAS platforms equipped with on-board AI accelerators running real-time computer-vision inference, train custom object-detection models, deploy them to the drone, and validate end-to-end autonomous decision-making in flight. This is the cutting edge of what threat actors can field today — and the only way to truly understand the AI-equipped sUAS threat is to build one yourself.

Built for technically-oriented bomb technicians, military EOD, and federal counter-UAS personnel who need to characterize, detect, exploit, and defeat AI-equipped drones in their operational environment.

Course Structure

Week 1 — Foundational Drone Build (Days 1–5). Same as our 1-Week Construction Build Class: electronics refresher, improvised initiation systems, frame and motor and ESC and FC and RX and VTX and camera build, payload integration, range day.

Week 2 — Edge AI Build, Train, Deploy.

• Day 6 — Edge AI Hardware Foundations. Install on-board AI accelerator, drivers, runtime. Boot, system update, camera-input pipeline.
• Day 7 — Pre-Trained Vision Models. Real-time object detection on the drone using open-source pre-trained models. Latency profiling, model optimization, calibration to specific target classes.
• Day 8 — Custom Model Training. Lab-based training of custom object-detection models on collected imagery. Annotation, dataset construction, training run, validation, deployment to the drone.
• Day 9 — On-Drone Autonomous Decision-Making. Behavior trees and decision logic linking AI inference to flight-controller actions: detect → classify → engage / report / loiter. Telemetry pipeline back to ground station for real-time AI output.
• Day 10 — Capstone + Range Day. Full autonomous mission with on-drone AI decision-making: students fly a mission, the drone detects targets and makes engagement decisions in flight, telemetry is captured, post-mission analysis with the ground-station log.

What’s Included

• 10 complete Week-1 student build kits (per the 1-Week Construction Build Class)
• 10 AI accelerator upgrade kits — on-board AI co-processor, high-resolution camera with auto-focus, 1080p digital FPV VTX + camera, integrated open-source flight controller and GPS combo board, 915 MHz telemetry radio set
• 4 shared ML model-training workstations with GPU (one per pair of student pairs)
• 5 sets of 1080p digital FPV goggles (1 per pair of students)
• Custom annotated training-data set (sUAS-relevant imagery) for model-training exercises
• Forensics tool kit (USB/UART/SWD adapters, SD-card recovery hardware, JTAG)
• 10 professional-grade bench stations — stays at your unit after the course:
  • Temperature-controlled soldering station (60W)
  • Hot-air rework station
  • Bench multimeter (auto-range)
  • Variable bench PSU (0–30 V, 0–5 A)
  • Pocket scope / logic analyzer
  • ESD mat, helping-hands, tweezers, cutters, consumables
• Custom-trained model artifacts (theirs to keep) + training-pipeline scripts
• AI ethics and acceptable-use briefing built into Day 10
• Printed lab guides + curriculum + STL files + ground-control mission profiles
• Lead instructor × 2 weeks, travel + lodging + per-diem to your unit (CONUS, OCONUS, or APO)
• Two-instructor upgrade available at +$10,000 for a 5:1 student ratio
• APO shipping consolidated through Improvised Electronics San Diego
• Post-course AAR + digital training reference package

Logistics

• Class size: 10 students (12 max with 2nd instructor at additional cost)
• Lead time: 10 weeks from PO to class start (longer than other courses due to AI hardware sourcing and custom dataset prep)
• Delivery: MTT to your unit, or CONUS-hosted at IE San Diego
• GSA Contract 47QRAA24D0055 | UEI NX6AA1JY7UF4 | NET30 / PO accepted