The project presented in the article uses:

• The right arm of the Inmoov, configuring the fingers in a non-linked double tendon arrangement.
• Inertial Measurement Units to determine the finger positions.
• Servo motors as actuators for the fingers.
• An interface for message exchange using the MQTT protocol to receive voice commands from a web application.
• Finger movements are controlled by an optimal PI controller.

To achieve the above, the ESP32 microcontroller is used, applying parallel multitasking programming techniques. The ESP32 folder contains the files that need to be installed on the microcontroller, where the following information needs to be updated:

• mpu6050Functions.cpp [Lines 73-112]: Offsets of the IMUs for each finger.
• mqttFunctions.cpp [Lines 17-20]: Subscriber and publisher topics, as well as username and password for the MQTT server.
  • The onMqttMessage* function [Line 123]: This section defines the codes for the commands that the prototype can execute.
    • a: Sending the current system status.
    • b: Sending the current values of the variables.
    • c: Sending the current PID gains.
    • g: Activating the grip mode.
    • h: Deactivating the grip mode.
    • s: Receiving the setpoint for each finger.
    • p: Receiving the new gains for the PID controller of each finger.
• mqttFunctions.h [Lines 13-17]: WiFi name and password. MQTT server name and port.
• utils.cpp, utilsSetup() function: Configuration of the gains for the PID of each finger.
• utils.h [Lines 4]: Pin number of the force sensor.

The web application developed for this article, through which voice commands can be sent to the hand via the MQTT protocol, is developed in Java using the Spring Framework. The WebApp folder contains the necessary files to import the application as a Spring/Maven project in the STS IDE.