Multi-step highway traffic flow prediction is crucial for intelligent transportation systems, and existing works have made significant advancements in this field. However, the physical structure, including path, distance, and node degree, is critical information in traffic networks and is often overlooked when encoding spatiotemporal dependencies. Meanwhile, the problem of prediction error propagation in multi-step flow forecasting is challenging to mitigate and can significantly impact overall forecasting performance. Moreover, traffic flows between toll and gantry stations in the highway network exhibit notable differences, leading to heterogeneous flow distributions. To overcome the above issues, a novel multi-task spatiotemporal network for highway traffic flow prediction (MT-STNet) is proposed, consisting of the encoder-decoder structure, a generative inference system, and multi-task learning. The spatiotemporal block with physical transformation is developed to construct both the encoder and decoder, integrating the physical structure information into modeling the highway network’s spatiotemporal dependencies. Additionally, the generative inference architecture is designed to extract the correlation between the historical-and target-sequences to generate the target hidden representations rather than a dynamic decoding way, avoiding multi-step prediction error propagation. Furthermore, because of traffic flow heterogeneity in the highway network, multi-task learning divides highway traffic flow prediction into three tasks, sharing the underlying traffic network and knowledge learned, thereby enhancing the prediction performance of each subtask. The evaluation experiments used monitoring data from a highway in Yinchuan City, Ningxia Province, China. The experimental results demonstrate that the performance of our proposed prediction model is better than that of the baseline methods.

Predicting how long a trip will take may allow travelers plan ahead, save money, and avoid traffic congestion. The journey time estimation model should take into account three crucial factors: (1) individual travel preference, (2) dynamic spatio-temporal correlations, and (3) the association between long-term speed forecast and travel time estimate. In order to overcome these challenges, this study proposes a unique parallel architecture called the multi-task spatio-temporal attention network (MT-STAN) to estimate journey times. To extract the dynamic spatio-temporal correlations of the road network, we first develop a traffic speed prediction model based on spatio-temporal block and bridge transformer networks, combining the road, timestamp, and traffic speed information into hidden states. Second, we offer a personalized model for estimating journey times that makes use of cross-network, holistic attention, and semantic transformer. In this approach, travel preferences extraction through cross-network, holistic attention permits correlations between the dynamic road network’s hidden states and individual journey characteristics, which are subsequently transformed into global semantics by the semantic transformer; preferences and semantics are integrated during the estimate phase. Finally, a multi-task learning component is included, which combines both traffic speed prediction and individual journey time estimate, via the sharing of underlying network parameters and the improvement of the contextual semantic knowledge of the latter job. Evaluation experiments are carried out using a highway dataset collected in Yinchuan City, Ningxia Province, China. The proposed prediction model outperforms state-of-the-art baseline approaches in experiments.