Bike-sharing is a new low-carbon and environment-friendly mode of public transport based on the “sharing economy”. Since 2017, the bike-sharing market has boomed in China’s major cities. Bikes equipped with GPS transmitters are docked along sidewalks that can be easily accessed through smartphone apps. However, this new form of transport has also led to problems, such as illegal parking, vandalism, and theft, each of which presents a major administrative challenge. Further, imbalances in user demand and bike availability need to be overcome to ensure a convenient, flexible service for customers. Hence, predicting a cyclist’s destination could be of great importance to shared-bike operators. In this paper, we propose an innovative deep learning model to predict the most probable destination for each user. The model, called destination prediction network based on spatiotemporal data (DPNst), comprises three steps. First, the data is preprocessed and a pool of likely candidate destinations is generated based on frequent item mining. This candidate set is then used to build the DPNst model: a long short-term memory network learns the user’s behavior; a convolutional neural network learns the spatial relationships between the origin and the candidate destinations; and a fully connected neural network learns the external features. In the final step, DPNst dynamically aggregates the output of the three neural networks based on the given data and generates the predictions. In a series of experiments on real-world stationless bike-sharing data, DPNst returned an F1 score of 42.71% and demonstrated better performance overall than the compared baselines.
Based on a heuristic optimization algorithm, this paper proposes a new algorithm named trajectory-planning beetle swarm optimization (TPBSO) algorithm for solving trajectory planning of robots, especially robot manipulators. Firstly, two specific manipulator trajectory planning problems are presented as the practical application of the algorithm, which are point-to-point planning and fixed-geometric-path planning. Then, in order to verify the effectiveness of the algorithm, this paper develops a control model and conducts numerical experiments on two planning tasks. Moreover, it compares with existing algorithms to show the superiority of our proposed algorithm. Finally, the results of numerical comparisons show that algorithm has a relatively faster computational speed and better control performance without increasing computational complexity.
Accurate and timely short-term traffic prediction is important for Intelligent Transportation System (ITS) to solve the traffic problem. This paper presents a hybrid model called SpAE-LSTM. This model considers the temporal and spatial features of traffic flow and it consists of sparse autoencoder and long short-term memory (LSTM) network based on memory units. Sparse autoencoder extracts the spatial features within the spatial-temporal matrix via full connected layers. It cooperates with the LSTM network to capture the spatial-temporal features of traffic flow evolution and make prediction. To validate the performance of the SpAE-LSTM, we implement it on the real-world traffic data from Qingyang District of Chengdu city, China, and compare it with advanced traffic prediction models, such as models only based on LSTM or SAE. The results demonstrate that the proposed model reduces the mean absolute percent error by more than 15%. The robustness of the proposed model is also validated and the mean absolute percent error on more than 86% road segments is below 20%. This research provides strong evidence suggesting that the proposed SpAE-LSTM effectively captures the spatial-temporal features of the traffic flow and achieves promising results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.