Perpustakaan Digital ITB

Conventional handheld UAV remote controllers are widely used because they provide familiar, repeatable, and functionally complete joystick-based control. However, their size, storage requirements, protruding components, and two-handed operation can limit rapid deployment in confined or time-critical environments. This research develops and evaluates a glove-based wearable remote-control system for multirotor UAV operation as an alternative human-control interface. The controller combines IMU-based hand-orientation input, physical throttle control, deliberate arming input, auxiliary controls, safety gating, and RC channel transmission using an ExpressLRS communication link with Crossfire Serial Protocol-formatted channel data. The research was conducted through an iterative design, manufacturing, and validation process. The prototype was evaluated through bench functionality testing, line-of-sight and FPV flight testing, simplified payload mission testing, operational demonstration, flight-log analysis, and respondent-based user testing. The user evaluation involved 10 respondents, consisting of 8 novice users and 2 experienced UAV pilots. The evaluation used setup time, task completion time, error and safety-event observation, usability response, NASA-TLX-inspired workload rating, controller preference, and qualitative feedback. Technical validation showed that the wearable controller could generate valid roll, pitch, yaw, throttle, arming, and auxiliary RC channel outputs. Flight-log analysis verified command-output consistency between the wearable command and the flight-controller RC input trend, while the actual UAV motion remained dependent on the flight-controller mode, tuning, and aircraft dynamics. The results showed that the wearable controller provided a practical advantage in deployment speed and initial task execution under the tested conditions. All novice respondents completed setup faster in the wearable condition, with the average setup time decreasing from approximately 2 minutes and 9 seconds to approximately 53 seconds. The average novice task completion time also decreased from approximately 3 minutes and 32 seconds to approximately 3 minutes and 13 seconds. However, respondent feedback and testing results showed that the prototype still requires improvement in command sensitivity tuning, accidental input prevention, run-gesture reliability, fatigue reduction, wearable fit, telemetry or display feedback, and long-term practicality.