Applications and Implications, Project Development

What will it do?

My final project will be a Wearable Robotic Arm Controller. It will consist of a smart glove that works as a wireless controller for a remote robotic arm. The glove will read the user's finger flexion movements and hand orientation in real time. This data will be processed, filtered to remove noise, and wirelessly transmitted to the base of the robotic arm, which will replicate the exact same movements using servomotors and a stepper motor.

Who's done what beforehand?

Gesture gloves and robotic arms are nothing new; they've been around for years. In the commercial sector, devices like the 5DT Data Glove and the CyberGlove exist, primarily used in research and simulation. And if we look at open-source projects, projects like InMoov and BCN3D Moveo have demonstrated that it's possible to build a functional robotic arm with 3D-printed parts and inexpensive electronic components.

What will I design?

3D Design

I will design the structural parts of the robotic arm and the ergonomic supports of the handheld device: the arm band that will hold the plate and the finger supports that will hold the sensors.

Electronics Design

I will design two custom modular boards. One transmitter board for the glove and one receiver/power controller board for the arm.

Programming

The C++ firmware will integrate signal filtering, a complementary filter for the MPU6050, and an EMA/median filter for the flex sensors, along with kinematic control logic.

What sources will I use?

Throughout the process, I will consult a lot of documentation and community resources. For the wireless part, I will go straight to Espressif's ESP-NOW documentation and Seeed Studio's XIAO ESP32-C3 datasheet. For the IMU, I will use Jeff Rowberg's MPU6050 library and its documentation. For the mechanical design, I will look at some open-source robotic arm models for inspiration on the joints and gripper before designing my own. The printed circuit boards will be designed in Altium, so I will use its libraries extensively. And for oversampling, median filtering, and EMA smoothing, I will research Arduino forums and some GitHub repositories to find out what actually works.

What materials and components will be used?

Below is a breakdown of every component that will be used in the project, where I will source it, and its approximate cost in US dollars. The original prices were in Mexican pesos, converted at an exchange rate of about 17.3 MXN per USD, so the figures are estimates that will shift with the exchange rate and with seller, stock, and shipping. I will mostly buy from Mercado Libre and Amazon México, with the fabric coming from Parisina and a few items from specialized electronics shops.

Electronics and Mechanical

Screws and Nuts

All of the screws below will be used together with their matching nuts. I will buy them from a local hardware store (tornillería), which works out much cheaper than buying them individually online.

What parts and systems will be made?

This system will consist of two physical parts. The glove will use a XIAO ESP32-C3, three flex sensors, an MPU6050, and a PCB that will integrate the power and signals. The arm will use a XIAO ESP32-C6 as a receiver, a NEMA stepper motor with its controller for base rotation, several PWM servos for the joints and gripper, an XL4016 to reduce the voltage from 12V to 6V for the servos, and another PCB that will manage all the wiring. All the structural parts on both sides will be 3D printed.

What processes will be used?

I will use 3D printing for all the mechanical parts. The PCBs will be designed in Altium and then fabricated. After that, there will be a lot of soldering, crimping connectors, and assembling everything into the printed enclosures. As for the software, I will write the firmware in C++ using the Arduino environment, including sensor calibration, filtering algorithms, and ESP-NOW communication.

What questions need to be answered?

I will need to sort out some questions along the way. What will the ESP-NOW's indoor range be before the signal becomes unstable? Will the latency be low enough for an instantaneous response? What filter combination will give me a smooth signal without a sluggish response? How will I calibrate the flex sensors for consistent readings? Will the stepper motor and servos be able to move the arm without getting stuck under the weight? How will I split the power between the 12V and 6V rails to avoid voltage drops or dips when multiple motors are moving simultaneously? And how much weight will the gripper handle before the gears or servo fail?

How will it be evaluated?

The arm must react to my hand quickly enough to feel like it's in real time, and it must precisely copy my movements, without any shudder or random motion. The wireless connection must remain stable throughout the entire demonstration without interruption. Furthermore, the structure must be solid, the printed circuit boards must function correctly without requiring any modifications, and the glove and arm must feel like a complete system, not two separate experiments. The real test will be performing a proper demonstration: rotating the wrist, opening and closing the fingers, and picking up and placing an object, using only the glove.