MODULE M9 · 10.0 HOURS
Performance Testing and Verification
LEARNING OBJECTIVES
Module objectives
- Design quantitative test metrics to verify the robot hand's precision and repeatability
- Evaluate the stability of the grasping force control algorithm utilizing FSR sensor data
- Analyze the error between DYNAMIXEL actuator feedback data and actual physical motion
- Learn procedures for verifying mechanical defects and the durability of the tendon drive mechanism
Performance Testing and Verification Theory
Robot hand performance verification is the process of confirming the consistency between design specifications and actual physical behavior [S1]. Key metrics include the following:
1. Positioning and Grasping Precision
Repeatability refers to the range of error in the position reached by the robot hand when performing the same command. While the XM430-W350-T actuator provides precise position feedback via its internal encoder [S11], the final position of the fingertip incurs errors due to tendon elongation and friction. Dyneema tendons have very low elongation (less than 1%), which is advantageous for ensuring repeatability [S16].
2. Force Control and FSR Sensor Signal Processing
The FSR 402 sensor is characterized by a decrease in resistance as the applied force increases [S12]. This is configured with a 10 kΩ resistor in a voltage divider circuit and measured with the OpenCR’s 12-bit ADC [S13, S26]. Since sensor data is noisy, a Moving Average Filter must be applied to form a stable grasping force feedback loop.
3. Overcurrent Protection and Power Stability
The system uses 3 independent 12 V power branches [S15]. Each branch is protected by a 10 A ATOF fuse [S25], and distribution must be managed so that the sum of the actuator peak currents does not exceed the protection rating. This must be verified by coordinating protection via the time-current curves provided by the manufacturer.
WORKED EXAMPLES
Worked examples
- Example 1: OpenCR ADC voltage calculation. When the FSR resistance is 10 kΩ and the series resistor is 10 kΩ, the 3.3 V voltage divider output is $V_{out} = 3.3 * (10k / (10k + 10k)) = 1.65 V. This is suitable for the 12-bit ADC range [S13, S26].
- Example 2: Fuse protection coordination. When 4 actuators are in a stall state, the sum of currents is 9.2 A [S11]. Since the cold resistance of a 10 A fuse is 7.7 mΩ [S25], the voltage drop during normal operation is approximately 0.07 V, which can be neglected; however, for overcurrent, refer to the fuse manufacturer's time-current curve for the exact response.
LAB PROTOCOL
Robot Hand Integrated Functional Testing
- 1
With each power branch physically disconnected, measure the output of the 3 adapters in DC voltage mode to confirm it reads 12 V.
- 2
Fix the robot hand in a safety jig and connect the controller (OpenCR) to a PC to release the actuator torque to 0.
- 3
Record the ADC data changes while manually applying pressure to the FSR sensor of each finger.
- 4
Repeat the maximum range of motion (ROM) for each finger 5 times in a no-load state to check for tendon interference.
- 5
After the test is complete, you must unplug the 3 power adapters from the wall outlet and check for residual voltage.
- Always wear safety glasses during testing.
- Do not place your hands within the range of motion while power is applied.
- If abnormal heat, odor, or smoke is detected, do not approach; evacuate after cutting the supply power to the 3 adapters at the pre-designated building distribution panel breaker or an authorized upstream master disconnect outside the hazardous zone. If there is no accessible upstream disconnection means outside the zone, do not energize the system. Torque release is not a substitute for power disconnection. Maintenance/access must only be performed after a planned shutdown, physical disconnection, and confirmation by de-energized measurement.
- Do not touch the system without measuring voltage. Verifying DC is less than 1 V is mandatory.
Lab deliverables
- Fingertip grasping force sensor calibration records
- Repeatability precision measurement data
- Load current measurements per power branch
ASSIGNMENT
Robot Hand Performance Analysis Final Report
Deliverables
Rubric
- Appropriateness of the Signal-to-Noise Ratio (SNR) analysis of sensor data
- Quantification of precision in repeatability tests
- Theoretical reflection on whether the protection design (fuses) satisfies the intended system protection
- Comparison of design specifications and actual built performance metrics
KNOWLEDGE CHECK
Knowledge check
FIELD CHECK
Completion criteria
- Submitted performance test result report and obtained at least 70 points
- Complied with safety guidelines in all Lab steps and confirmed physical power disconnection
- Confirmed implementation of the sensor data filtering function in the control code
MODULE SOURCES
Module sources
- Introduction to Robotics: Mechanics and Control - John J ... John J.Craig-Introduction to Robotics Mechanics and Control ... Introduction to robotics : mechanics and control : Craig ... Introduction to Robotics: Mechanics and Control (3rd Edition) Introduction to robotics : mechanics and control : Craig ... Introduction to Robotics: Mechanics and Control - John J ...books.google.com · textbook
- XM430-W350 e-Manualemanual.robotis.com · datasheet
- D-Pro 1.5 mmliros.com · technical_documentation
- FSR Model 402interlinkelectronics.com · datasheet
- OpenCR 1.0emanual.robotis.com · technical_documentation
- 10K Ohm 5 Percent 1/4W Through-Hole Resistorsadafruit.com · technical_documentation
- GST160A12-R7B Enclosed Desktop Power Adaptermeanwell.com · datasheet
- ATOF 287 Series 10 A Blade Fuselittelfuse.com · datasheet