GESTUROBO: REVOLUTIONIZING EDUCATIONAL ROBOTICS WITH WIRELESS TINYML CONTROL - DIIID 2026

[00:00:03][music] [music] >> Hello honorable judge. We [music] are from Polytechnic Caltex Rio.

[00:00:21]In educational robotic in laboratories, 6 DOF robotic arm are frequently used [music] to help student understand mechatronic.

[00:00:28]However, conventional control method like wired potentiometer [music] present significant limitation.

[00:00:35]They restrict user mobility, are highly susceptible to cable [music] fatigue, and are non-intuitive for novice user because they require sequential manual adjustment.

[00:00:46]>> To address this issue, we designed Gesturobo, a wireless gesture-based human-machine interface.

[00:00:53]Our main objective is to eliminate user mobility restrictions and the cable fatigue associated with conventional tethered arms.

[00:01:06]We aim to provide a more intuitive, natural, and efficient control solution for beginners in educational robotics.

[00:01:18]>> The hardware architecture [music] is perfectly practical. We designed a custom 3D print hand held transmitter and receiver [music] modules. Inside this compact case, we use ESP32-C3 super mini microcontroller connected to an [music] MPU-6050 IMU sensor to capture hand orientation and motion.

[00:01:35]>> [music] >> As you can see, the building screen instantly display a recognized gesture.

[00:01:40]This data is transmitted wirelessly in real time to the receiver module to seamlessly actuate to servo motor on the 6 DOF robotic arms.

[00:01:50]>> [music] >> To keep the system lightweight and suitable for embedded hardware, we implemented a threshold-based classification method. [music] The raw sensor readings are stabilized using a basic smoothing.

[00:02:03]Then, the values are compared with predefined threshold [music] range.

[00:02:07]Our system can accurately recognize 10 specific motion commands [music] such as left, right, forward, backward, pitch, and pan, translating them commands [music] into robotic command signal.

[00:02:21]Let's see the live demonstration. As you can see, the response [music] is instantaneous and wireless. Based on our experimental trials, the Gestorover achieved a gesture recognition accuracy of 95.90% [music] across the 10 predefined commands.

[00:02:37]Furthermore, the system recorded a 100% success rate in gesture to cell phone mapping. Practically, the wireless control [music] reduced task completion time by 76.2% making it 4.2 >> [music] >> times faster than the conventional potentiometer based interface.

[00:02:54]In conclusion, [music] the Gestorover system utilizing an ESP32-C3 and MPU6050 [music] is a highly effective and reliable alternative to traditional tethered interface. We successfully [music] eliminate cable limitation, enhance user ergonomics, and significantly speed up operation times.

[00:03:15]Gestorover is an ideal intuitive [music] solution for the future educational robotic. Thank you for your time and keep innovating.

[00:03:23]>> [music]