Why Your FTC Autonomous Breaks on the Tournament Field (and How to Fix It)
Your autonomous routine fails at competition because dead reckoning relies on physical constants that do not exist: mat friction, tile density, and battery voltage vary dramatically between your workshop and the arena.
When an autonomous routine executes ten consecutive flawless runs in your classroom only to crash into the perimeter wall during Match 1, the code is rarely the culprit. The underlying physics changed. Pure dead reckoning—whether timed runs or standard wheel encoder counts—assumes an immutable physical environment. In reality, competition fields introduce three silent variables that destroy open-loop navigation: foam tile compression, surface friction discrepancies, and battery discharge curves.
1. The Mat Problem: Density, Age, and Compression
Official FTC interlocking foam tiles (traditionally SoftTiles) are not uniform across venues. A practice field tile in a school robotics lab that has endured two seasons of foot traffic and midday sun through a classroom window has compressed, hardened, and lost surface texture. In contrast, the tournament field may feature brand-new tiles out of the box, or tiles that have been stored tightly rolled in a provincial venue's humid storeroom.
This variation affects your robot in two distinct ways:
- Effective Wheel Radius: A heavier robot sinks deeper into softer foam. As the foam compresses around the wheel tread, the effective rolling radius shrinks. An encoder counting 1,000 ticks expects the robot to have travelled a specific linear distance based on nominal wheel diameter; on softer tiles, the actual linear distance decreases per revolution.
- Turn Scrub Resistance: Mecanum and standard skid-steer drivetrains experience vastly different lateral resistance depending on whether the foam gives way or resists sideways scrubbing. On grippier or softer foam, turns under-rotate; on worn, dusty foam, they over-rotate.
2. Voltage Sag and Acceleration Profiles
A fresh REV Robotics 12V Slim Battery directly off a smart charger reads approximately 13.8V to 14.2V open-circuit. After sitting in the staging queue for fifteen minutes, or after a heavy autonomous sequence where four drive motors and an intake draw peak stall currents, that internal voltage sags.
If your autonomous code relies on running motors at fixed power levels for set durations, the distance travelled is a direct function of the voltage curve over time. Even with PID velocity control on the drive motors, high-acceleration setpoints on a slightly drained battery will cause wheel slip at the start of a motion profile. The encoder records the wheel turning, but the robot has not moved the corresponding distance on the carpet.
3. Morning Pit Calibration Habits
Top-tier teams do not assume the field matches their workshop. They build a strict thirty-minute calibration routine into their morning schedule between registration and team briefing.
| Check-in Task | Target Parameter | Field Adjustment Method |
|---|---|---|
| Field Tile Scrub Test | Turning friction & sinkage | Drive a 1,000mm straight run on practice field; measure actual linear displacement with a physical tape. |
| IMU Drift Verification | Yaw stability | Sample REV Control Hub or external IMU for 60 seconds stationary to check gyro offset baseline. |
| Autonomous Battery Gate | Consistent voltage profile | Establish a strict rule: competition autonomous batteries must measure between 13.4V and 13.8V under resting load. |
| Webcam Exposure Locking | Apriltag / Vision reliability | Lock exposure and gain manually under arena halogen/LED lighting rather than relying on auto-exposure. |
To eliminate dead reckoning errors entirely, transition away from pure wheel encoders toward three-wheel dead-wheel odometry pods paired with an onboard IMU (Inertial Measurement Unit), or anchor your critical scoring positions to field landmarks using AprilTags or distance sensors. If you are retrofitting your drivetrain for the regional tournament circuit, our hardware selection at Sheen Store includes local-stock odometry components, tracking sensors, and REV-compatible cables to prevent signal dropouts.
4. The Software Safeguard: Tolerance and Resets
Finally, inspect your autonomous state machine. Code designed without timeouts or re-alignment steps is fragile by definition. If your intake mechanism relies on being within 5mm of a spike mark, build a physical squaring motion against the perimeter wall or utilize colour and distance sensors to dynamically re-zero the robot's coordinate frame before attempting a critical scoring cycle.
For teams looking to refine their programming architectures, motion profiling algorithms, and sensor fusion techniques ahead of provincial qualifiers, explore our dedicated competition coaching modules at Sheen FTC Support.
