Implements PID auto-tuning ROS2 node using relay feedback (Astrom-Hagglund) method. Service-triggered relay oscillation measures ultimate gain (Ku) and ultimate period (Tu), then computes Ziegler-Nichols PD gains. Safety abort on >25deg tilt. Features: - Service /saltybot/autotune_pid (std_srvs/Trigger) starts tuning - Relay oscillation method for accurate gain measurement - Measures Ku (ultimate gain) and Tu (ultimate period) - Computes Z-N PD gains: Kp=0.6*Ku, Kd=0.075*Ku*Tu - Real-time safety abort >25° tilt angle - JSON telemetry on /saltybot/autotune_info - Relay commands on /saltybot/autotune_cmd_vel Tuning Process: 1. Settle phase: zero command, allow oscillations to die 2. Relay oscillation: apply +/-relay_magnitude commands 3. Measure peaks: detect zero crossings, record extrema 4. Analysis: calculate Ku from peak amplitude, Tu from period 5. Gain computation: Ziegler-Nichols formulas 6. Publish results: Ku, Tu, Kp, Kd Safety Features: - IMU tilt monitoring (abort >25°) - Max tuning duration timeout - Configurable settle time and oscillation cycles Published Topics: - /saltybot/autotune_info (std_msgs/String) - JSON with Ku, Tu, Kp, Kd - /saltybot/autotune_cmd_vel (geometry_msgs/Twist) - Relay control Subscribed Topics: - /imu/data (sensor_msgs/Imu) - IMU tilt safety check - /saltybot/balance_feedback (std_msgs/Float32) - Balance feedback Package: saltybot_pid_autotune Entry point: pid_autotune_node Service: /saltybot/autotune_pid Tests: 20+ unit tests covering: - IMU tilt extraction - Relay oscillation analysis - Ku/Tu measurement - Ziegler-Nichols gain computation - Peak detection and averaging - Safety limits (tilt, timeout) - State machine transitions - JSON telemetry format Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Description
SaltyLab self-balancing bot firmware (STM32F722)
Languages
Python
67.1%
C
11.4%
JavaScript
9.2%
OpenSCAD
7.8%
HTML
1.5%
Other
2.9%