Exploring Trajectory Prediction through Machine Learning Methods
Keywords:
Long Short-Term Memory (LSTM), Sequence-to-Sequence (Seq2Seq)Abstract
This project presents a machine learningbased
approach for predicting the future trajectory
of users by analyzing their previous movement
patterns. Utilizing advanced algorithms such as
Long Short-Term Memory (LSTM) networks and
Sequence-to-Sequence (Seq2Seq) models, the
system aims to forecast the next location(s) of a
user based on historical GPS data. Accurate
location prediction has significant applications in
5G networks, where allocating the nearest cloud
resources to a user can drastically reduce latency
and enhance user experience. The Geolife dataset,
which consists of real-world GPS trajectories
including latitude, longitude, and user ID, was used
to train the model. Experimental results show that
the proposed method achieves high prediction
accuracy, with an LSTM-Seq2Seq model
outperforming traditional algorithms like GRU,
demonstrating lower mean squared error rates and
better trajectory forecasting performance.
Downloads
References
1) Hochreiter, S., & Schmidhuber, J. (1997). Long
short-term memory. Neural Computation, 9(8),
1735–1780.
2) Sutskever, I., Vinyals, O., & Le, Q. V. (2014).
Sequence to sequence learning with neural
networks. Advances in Neural Information
Processing Systems, 27.
3) Zheng, Y., Chen, Y., Li, Q., Xie, X., & Ma, W.
Y. (2010). Understanding transportation modes
based on GPS data for web applications. ACM
Transactions on the Web, 4(1), 1–36.
4) Liu, Y., Zhang, J., Zheng, Y., & Xie, X. (2017).
Urban computing: Using big data to understand
and improve urban dynamics. Frontiers of
Computer Science, 11(2), 167–177.
5) Alahi, A., Goel, K., Ramanathan, V.,
Robicquet, A., Fei-Fei, L., & Savarese, S.
(2016). Social LSTM: Human trajectory
prediction in crowded spaces. CVPR.
6) Cho, K., et al. (2014). Learning phrase
representations using RNN encoder-decoder for
statistical machine translation. arXiv preprint
arXiv:1406.1078.
7) Wang, Y., Li, Z., & Wu, X. (2018). Deep
learning-based intelligent trajectory prediction.
IEEE Access, 6, 51258–51266.
8) Yao, H., Tang, X., Wei, H., Zheng, G., & Li, Z.
(2018). Revisiting spatial-temporal similarity:
A deep learning framework for traffic
prediction. AAAI Conference on Artificial
Intelligence.
9) Zhang, J., Zheng, Y., & Qi, D. (2017). Deep
spatio-temporal residual networks for citywide
crowd flows prediction. AAAI.
10) Xu, C., He, Y., & Song, R. (2020). Trajectory
prediction using deep learning methods: A
review. IEEE Transactions on Intelligent
Transportation Systems, 22(3), 1341–1359.
