TY - GEN
T1 - Leader Follower Navigation System based on Pedestrian Dead Reckoning for Mobile Robot Navigation
AU - Farih, Muhammad
AU - Sahal, Mochammad
AU - Rusdhianto Effendi, A. K.
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/8
Y1 - 2019/8
N2 - Navigation systems are the first things to have for all types of unmanned vehicles whether it is land, water, or air vehicles. Navigation systems are divided into independent and leader-follower. Today, Global Positioning System (GPS) plays an important role for vehicle positioning. Nevertheless, GPS has a weakness that the signals will be distorted when the GPS receivers are placed in indoor environments. In this research, leader-follower navigation systems based on pedestrian dead reckoning are designed with wheeled mobile robot as the follower. Artificial Neural Network (ANN) is used as stride length model and the leader heading is obtained from magnetometer. Both data are sent to the follower through Wi-Fi. For follower system, linear model is used as robot distance model and follower heading is obtained from its magnetometer. The test shows that the best model for stride length prediction is ANN model with one hidden layer and four neuron units which has 4.65 cm training error and 5.04 cm testing error using 614 stride examples. As for robot distance model shows that the error is 2.0956 cm using 25 data samples. Finally, the heading error is 39.262° which is tested from two magnetometers with 28 testing points.
AB - Navigation systems are the first things to have for all types of unmanned vehicles whether it is land, water, or air vehicles. Navigation systems are divided into independent and leader-follower. Today, Global Positioning System (GPS) plays an important role for vehicle positioning. Nevertheless, GPS has a weakness that the signals will be distorted when the GPS receivers are placed in indoor environments. In this research, leader-follower navigation systems based on pedestrian dead reckoning are designed with wheeled mobile robot as the follower. Artificial Neural Network (ANN) is used as stride length model and the leader heading is obtained from magnetometer. Both data are sent to the follower through Wi-Fi. For follower system, linear model is used as robot distance model and follower heading is obtained from its magnetometer. The test shows that the best model for stride length prediction is ANN model with one hidden layer and four neuron units which has 4.65 cm training error and 5.04 cm testing error using 614 stride examples. As for robot distance model shows that the error is 2.0956 cm using 25 data samples. Finally, the heading error is 39.262° which is tested from two magnetometers with 28 testing points.
KW - Artificial Neural Network
KW - Leader follower navigation systems
KW - Magnetometer
KW - Pedestrian Dead Reckoning
UR - http://www.scopus.com/inward/record.url?scp=85078457987&partnerID=8YFLogxK
U2 - 10.1109/ISITIA.2019.8937138
DO - 10.1109/ISITIA.2019.8937138
M3 - Conference contribution
AN - SCOPUS:85078457987
T3 - Proceedings - 2019 International Seminar on Intelligent Technology and Its Application, ISITIA 2019
SP - 96
EP - 101
BT - Proceedings - 2019 International Seminar on Intelligent Technology and Its Application, ISITIA 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2019 International Seminar on Intelligent Technology and Its Application, ISITIA 2019
Y2 - 28 August 2019 through 29 August 2019
ER -