Year: 2026
Type: article
Abstract: Teleoperation systems allow operators to control robots remotely, yet low-cost motion-capture interfaces are typically limited by short-range wireless links. This study implements and evaluates a nine-actuator anthropomorphic robot arm teleoperation system driven by a five-flex-sensor glove and an MPU-6050 IMU, using Raspberry Pi Zero 2 W processors on both ends and a point-to-point 915 MHz LoRa SX1278 link (SF9/BW 125 kHz/CR 4/5, 48-byte payload). Flex sensor signals are digitized by a 12-bit MCP3208 ADC over SPI, smoothed by a moving-average filter (N=5), calibrated using a two-point method, and mapped to servo PWM via a PCA9685 with per-joint discrete PID control, velocity limiting, and a dead-zone. Tests covered motion-imitation accuracy (4 scenarios × 30 repetitions), end-to-end latency (200 iterations at 10/100/200 m), LOS/NLOS range up to 200 m (500 packets per point at SF7/SF9/SF12), and power consumption. Results show an average accuracy of 92.3% (90.4–94.6% per joint), mean latency from 113.9 ms (10 m) to 119.6 ms (200 m) with 95th percentiles below 165 ms, an SF9 PDR of 97.8% at 200 m LOS, path-loss exponents of n = 2.23 (LOS) and 3.08 (NLOS), and a transmitter power draw of 1.32 W during active transmission. All performance targets were met, and one-way ANOVA confirmed significant latency differences across SF configurations (p < 0.001). The system is a feasible low-cost, portable teleoperation solution with a range of hundreds of meters.
Institution: Politeknik Negeri Manado
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7218255653