TY - GEN
T1 - Multi-Model Bus Arrival Time Prediction using Real-Time Online Information
AU - Yang, Zihao
AU - Lam, Chan Tong
AU - Ng, Benjamin K.
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Precise bus arrival time prediction can attract more passengers to take public transportation. In this paper, we propose a multi-model bus arrival time prediction approach using real-time online information, by decomposing the travel time into the sum of dwell time and link time, which are predicted separately and then summed to obtain the predicted trip time between adjacent bus stops. Four prediction models, namely Simple Moving Average (SMA), Artificial Neural Network (ANN), Long Short Term Memory (LSTM) and Hybrid Model (SMA and LSTM), are used to evaluate and compare the performance of the proposed multi-model prediction approach. Using the proposed hybrid multi-model, it was found that the average MAPE% among SMA, ANN, LSTM and Hybrid Model are 33.56%, 32.15%, 26.76% and 23.45%, respectively. The proposed hybrid multi-model with SMA and LSTM gives the best performance, with a MAPE improvement of about 3.3%.
AB - Precise bus arrival time prediction can attract more passengers to take public transportation. In this paper, we propose a multi-model bus arrival time prediction approach using real-time online information, by decomposing the travel time into the sum of dwell time and link time, which are predicted separately and then summed to obtain the predicted trip time between adjacent bus stops. Four prediction models, namely Simple Moving Average (SMA), Artificial Neural Network (ANN), Long Short Term Memory (LSTM) and Hybrid Model (SMA and LSTM), are used to evaluate and compare the performance of the proposed multi-model prediction approach. Using the proposed hybrid multi-model, it was found that the average MAPE% among SMA, ANN, LSTM and Hybrid Model are 33.56%, 32.15%, 26.76% and 23.45%, respectively. The proposed hybrid multi-model with SMA and LSTM gives the best performance, with a MAPE improvement of about 3.3%.
KW - ANN
KW - Bus arrival time prediction
KW - Hybrid model
KW - LSTM
KW - Real-Time Online Information
KW - SMA
UR - http://www.scopus.com/inward/record.url?scp=85152291253&partnerID=8YFLogxK
U2 - 10.1109/ICCT56141.2022.10072901
DO - 10.1109/ICCT56141.2022.10072901
M3 - Conference contribution
AN - SCOPUS:85152291253
T3 - International Conference on Communication Technology Proceedings, ICCT
SP - 1918
EP - 1922
BT - 2022 IEEE 22nd International Conference on Communication Technology, ICCT 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 22nd IEEE International Conference on Communication Technology, ICCT 2022
Y2 - 11 November 2022 through 14 November 2022
ER -