Performance Benchmarks and Stability Testing

Control Loop Benchmarks

Target Specifications

Metric	Target	Method
Motor PWM update rate	50 Hz	LEDC timer measurement
Odometry publish rate	50 Hz	ROS2 topic Hz monitor
Ultrasonic scan cycle	10 Hz (2.5 Hz/sensor)	Timer ISR measurement
IMU data rate	100 Hz	I2C transaction timing
cmd_vel → motor latency	< 20 ms	GPIO toggle measurement
E-stop response time	< 5 ms	ISR + relay timing

Measurement Procedure

1. Control Loop Jitter

# On companion computer, monitor topic rates
ros2 topic hz /odom --window 100
ros2 topic hz /cmd_vel --window 100

# Expected output:
# /odom: average rate: 50.0 Hz, min: 19.5ms, max: 20.5ms, std dev: 0.3ms

2. End-to-End Latency

Hardware setup: Connect GPIO test pin to oscilloscope.

// In firmware, toggle test pin on cmd_vel receive and motor update
gpio_set_level(TEST_PIN, 1);  // On cmd_vel callback
// ... process ...
gpio_set_level(TEST_PIN, 0);  // After motor PWM set

Measure pulse width on oscilloscope. Target: < 20ms (1 control cycle).

3. E-Stop Response Time

Test procedure: 1. Robot moving at maximum speed (1.2 m/s) 2. Press E-stop button 3. Measure time from button press to motor current = 0

Equipment: Current probe on motor leads + E-stop GPIO signal.

Expected: < 5ms (single ISR latency + relay activation).

4. Wi-Fi Latency (DTLS)

# On client machine
python3 -c "
import socket, time, struct

sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
robot_ip = '192.168.1.100'

for i in range(100):
    t0 = time.monotonic_ns()
    sock.sendto(b'PING' + struct.pack('Q', t0), (robot_ip, 8888))
    data, addr = sock.recvfrom(256)
    rtt = (time.monotonic_ns() - t0) / 1e6
    print(f'RTT: {rtt:.2f} ms')
"

Target RTT: < 10ms on local network, < 50ms on congested Wi-Fi.

---

Stability Test Suite

Test 1: 24-Hour Continuous Operation

Purpose: Verify no memory leaks, task starvation, or watchdog resets.

Procedure: 1. Power on robot with all systems active 2. Run continuous cmd_vel at 0.5 m/s in square pattern 3. Monitor for 24 hours 4. Log: free heap, task stack watermarks, restart count

Pass criteria: - Zero unplanned restarts - Free heap drift < 1 KB/hour - No task stack overflow warnings - All sensor topics maintain rated Hz ±5%

Monitoring script:

#!/bin/bash
# Run on companion computer
LOG="stability_$(date +%Y%m%d_%H%M%S).csv"
echo "timestamp,heap_free,stack_min,odom_hz,temp_c" > "$LOG"

while true; do
    HEAP=$(ros2 topic echo /diagnostics --once | grep heap_free | awk '{print $2}')
    STACK=$(ros2 topic echo /diagnostics --once | grep stack_min | awk '{print $2}')
    ODOM_HZ=$(ros2 topic hz /odom --window 10 2>&1 | grep average | awk '{print $4}')
    TEMP=$(ros2 topic echo /thermal --once | grep temp | awk '{print $2}')
    echo "$(date +%s),$HEAP,$STACK,$ODOM_HZ,$TEMP" >> "$LOG"
    sleep 60
done

Test 2: Payload Hot-Plug Stress

Purpose: Verify reliable hot-plug detection and power sequencing.

Procedure: 1. Insert/remove payload module 100 times 2. Vary insertion speed (slow: 2s, fast: 0.1s) 3. Record success/failure for each cycle

Pass criteria: - 100% detection rate for properly-seated insertions - Zero false positives from partial insertions (debounce working) - Power sequencing timing within spec (5V before 12V, 10ms ramp) - No EEPROM read failures on valid modules

Test 3: Wi-Fi Reconnection

Purpose: Verify automatic recovery from Wi-Fi drops.

Procedure: 1. Establish DTLS connection 2. Disable AP for 5 seconds 3. Re-enable AP 4. Verify reconnection and command resumption

Pass criteria: - Robot stops within watchdog timeout (500ms) on disconnect - Reconnection within 10 seconds of AP availability - No stale commands executed after reconnection - Session keys renegotiated on reconnect

Test 4: Thermal Stress

Purpose: Verify thermal protection under sustained load.

Procedure: 1. Run all 4 motors at 100% duty cycle 2. Monitor INA219 power readings 3. Continue until thermal warning triggers

Pass criteria: - Warning triggers before junction temp estimate > 70°C - Shutdown triggers before > 85°C - Motors stop within 1 control cycle of shutdown - Recovery after cool-down is automatic

Test 5: Battery Depletion

Purpose: Verify graceful shutdown at low battery.

Procedure: 1. Run robot until battery reaches 10% SOC 2. Verify low-battery warning published 3. Continue until critical threshold (5%)

Pass criteria: - Low battery warning at 10% SOC - Motor power reduced at 8% SOC - Graceful shutdown initiated at 5% SOC - No data corruption or brown-out resets

Test 6: Concurrent Payload + Navigation

Purpose: Verify payload I2C traffic doesn't interfere with control.

Procedure: 1. Active navigation (Nav2) running 2. Payload performing continuous I2C transactions (100 bytes/s) 3. Run for 1 hour

Pass criteria: - Odometry jitter < 2ms (no I2C bus starvation) - IMU update rate maintained at 100Hz - Payload transactions complete without CRC errors - No navigation drift from delayed encoder reads

---

Benchmark Results Template

Robot Platform Performance Report
=================================
Date:       YYYY-MM-DD
FW Version: x.y.z
SKU:        [Basic|Standard|Pro|Industrial]
Tester:     [Name]

Motor PWM Rate:     ___ Hz (target: 50)
Odom Publish Rate:  ___ Hz (target: 50)
cmd_vel Latency:    ___ ms (target: <20)
E-Stop Response:    ___ ms (target: <5)
Wi-Fi RTT:          ___ ms (target: <10)
Free Heap (start):  ___ bytes
Free Heap (24h):    ___ bytes
Heap Drift:         ___ bytes/hr (target: <1024)
24h Restarts:       ___ (target: 0)
Hot-plug Success:   ___/100 (target: 100/100)
Thermal Warning @:  ___°C (target: <70)
Thermal Shutdown @: ___°C (target: <85)

RESULT: [PASS|FAIL]
Notes: