Automatic Drug Delivery

Project Objective

Develop an affordable, automated device for opening an intravenous feed line using the venipuncture process, providing an economical alternative to current vein visualization tools while maintaining accuracy and safety.

Key Features

Budget-friendly (under $1000)
Portable and compatible with the Geriatric IV Training Arm
Accurate vein visualization and injection point localization
Compact design for home use (1.5x1.5 ft surface area)
Fast operation time (completed the mission in 2 minutes, under the 10-minute specification)

Design Specifications

Machine requirements:
- Maximum budget of $1000
- Portable and able to operate on the Geriatric IV Training Arm
- Dedicated power supply
- Robust construction
- Must not damage anything with which it interacts, including but not limited to the training arm
Robot tasks:
- Pick any visible vein sufficient for the needle tip dimensions
- Inject fluid into the vein using a syringe and withdraw the needle at the end of the operation

Perception Subsystem

The RealSense D435i camera is used to detect the exposed veins on the training arm and localize an optimal point for inserting the needle tip of the syringe.

This subsystem plans the trajectory for the needle to reach the insertion point position. The planner uses inverse kinematics to calculate the motor rotations, and then sends these signals to the motors, which rotate to the specified location.

The control subsystem uses an Arduino MEGA microcontroller to support and integrate the stepper motors and actuators, bridging the mechanical design subsystem with the perception subsystem

The physical structure, made up of a rectangular aluminum 80/20 1010 T-Slotted Profile frame and a two-axis gantry, controls the X and Y coordinates of the syringe. An end effector holds the syringe and includes a servo motor for rotation about the Z-axis and a linear actuator for fluid injection.

Motion Planning Subsystem

Control Subsystem

Mechanical Design Subsystem

Testing and Performance Evaluation

During the public demonstration, the device successfully detected and selected the best target injection point, moved the needle tip to the selected target location, performed an injection, and returned the needle to its starting position within two minutes. Although the device achieved the specified 0.5 mm tolerance, the fixed z-axis led to some veins on the arm being unreachable.

Conclusions and Future Recommendations

While the device met most specifications, it did not successfully inject fluid into the veins. Further improvements such as adding a third axis in the gantry, using a NIR camera for vein detection, and enhancing the perception subsystem for arm detection are recommended to improve the device's performance and reliability.