Embedded Files
Sensor Package
For this project, accurate measurement of motion, force, and torque was essential. Our finalized testbed was designed to capture detailed data of a person unlocking a lock using high-resolution sensors. To explore the range and variability of required forces and torques, we intentionally modified several keys to increase task difficulty and simulate real-world inconsistencies.
Force and torque resolution requirements were determined through empirical testing with various lock and key combinations. A sampling frequency of 1000 Hz was specified by the project sponsor to ensure adequate temporal resolution. We selected the ATI Mini45 force/torque sensor—already available through Professor Delson’s lab—because it met our performance requirements, including resolution and sampling rate, without incurring additional costs or procurement delays. The ATI Mini45 sensor is used to measure the forces and torques exerted during the unlocking process, helping us quantify exactly how much effort is required to operate the lock under different conditions.
For motion tracking, we chose the OptiTrack V:120 Duo. This system was selected for its robustness to electromagnetic interference (EMI), stable 120 Hz sampling rate, portability, and cost-effectiveness. Most importantly, it delivers sub-millimeter accuracy, which satisfies our project's precision requirements. The OptiTrack system is used to track the position and orientation of the robot system’s end-effector, particularly the Stewart platform, to ensure precise control and monitoring of its movements during the unlocking task.
Sensor Handle: This Handheld device will be used to collect human demonstration data.
Robot Arm
The robotic system was designed to provide six degrees of freedom and replicate the motions of a human unlocking a lock. It operates through a four-step linear process, controlled by a Raspberry Pi 5 microcontroller unit (MCU) capable of both manual and autonomous control.
Step 1: Key Insertion
The first step involves translating the key into the lock. This is accomplished using a linear rail actuated via GPIO connections from the Raspberry Pi 5. The rail supports two functions: insert and retract, both requiring a user-specified displacement input. During insertion, the rail moves the end-effector (key) forward toward the lock based on the commanded displacement.
Step 2: Jiggle Motion
The second step simulates the "jiggling" motion often required to align internal lock components. This step utilizes the Stewart platform, which provides six degrees of freedom for precise positional and orientational movement. Communication between the Raspberry Pi 5 and the Stewart platform’s microcontroller occurs via UART serial. The Raspberry Pi 5 reads CSV data from the OptiTrack motion capture system—representing recorded human unlocking motions—and converts it into movement commands for the Stewart platform, allowing it to replicate the position and orientation changes necessary to perform the jiggle.
Step 3: Key Rotation (Servo Roll Motion)
In the third step, the robot rotates the key to unlock the lock. A Dynamixel servo motor, mounted on top of the Stewart platform, performs this 90-degree rotational motion. It is controlled through UART communication between the Raspberry Pi 5 and an OPENRB-150 board. The Raspberry Pi sends two parameters to the servo: profile velocity and target position. The angular velocity is calculated using the following conversion:
Step 4: Key Retraction
Finally, the linear rail is used again to retract the key from the lock. This step is the inverse of insertion and uses the retract function of the rail, also requiring a displacement input. The key is pulled away from the lock to complete the unlocking cycle.
Wiring & Power Setup for the Robot System: A power strip will power the three robotic components explained in this section, the computer, and the monitor. The three robotic components, monitor, keyboard, and mouse will be connected to the Raspberry Pi 5.
Final Test Bed Design
Final Test Bed Design
This test bed integrates both the sensor package and the robot arm into one system. Optitrack and ATI sensors can track motions for human demonstration and then the robot system will mimic those motions based on data collected.
Page updated
Report abuse