A deep learning based approach for trajectory estimation using geographically clustered data

Abstract

This study presents a novel approach to predict a complete source to destination trajectory of a vehicle using a partial trajectory query. The proposed architecture is scalable to extremely large-scale data with respect to the dense road network. A deep learning model Long Short Term Memory (LSTM) has been used for analyzing the temporal data and predicting the complete trajectory. To handle a large amount of data, clustering of similar trajectory data is used that helps in reducing the search space. The clusters based on geographical locations and temporal values are used for training different LSTM models. The proposed approach is compared with the other published work on the parameters as Average distance error and one step prediction accuracy The one-step prediction accuracy is as good as 81% and Distance error are .33 Km. Our proposed approach termed Clustered LSTM is outperforming in both the parameters when compared with other reported results. The proposed solution is a clustering-based predictive model that effectively contributes to accurately handle the large scale data. The outcome of this study leads to improvise the navigation systems, route prediction, traffic management, and location-based recommendation systems.

A Comprehensive Review on Deep Learning-Based Motion Planning and End-to-End Learning for Self-Driving Vehicle

“…LaneGCN Architecture proposed by [45] performance and contribute to the further development of SDV.Additionally, a DL-based methodology for trajectory prediction using regionally clustered data has been developed by Aditya Shrivastava et. al [46]. To estimate a moving object's trajectory based on its prior locations and timestamps, the suggested method used LSTM neural networks.…”

“…Chanyoung Jung [62], D V Prasad Mygapul [66] Nejad [41] Wei Tian [42] Izquierdo [43] Pin Lv [44] Bing zhou [45] Shrivastava [46] Qianxia cao [40] Z Bai [47] Song [48] Cui [49] Stefano Pini [37] Dan Wang [38] Grigorescu [50] Jeong[51] Leordeanu [52] Behavioral Planning…”

Deep Learning Algorithms for predicting the geographical locations of Pandemic Disease Patients from Global Positioning System (GPS) Trajectory Datasets

“…In this paper, six (6) due to the unavailability of cheap computer power, high-capacity micro-processor as well as large storage systems, the growth was limited. Subsequently, the birth of Internet of Things (IoT) which led to the advent of big data and the availability of cheap computer power opened doors for increased use of DL models such as deep neural networks.…”

HeGA

“…Effectiveness of Neighbor Selection Methods. To study the effectiveness of NS methods, including heuristic-based methods (i.e., full-coverage, elliptical [7], rectangle [10], 𝑘-NN [34]) and our ANS method, we compare the average ADE of HeGA under these NS methods. The results are shown in Table 4.…”

“…Existing methods [1,7,10,25,34] select the surrounding agents based on a fixed region (e.g., rectangular or elliptical region) or distance-based clustering (e.g., 𝑘-nearest neighbors). Then they use an encoder-decoder LSTM framework to predict the future trajectory of the target agent based on the historical movements of both the target agent and its surrounding agents.…”