Trajectory distributions: A new description of movement for trajectory prediction
Abstract Trajectory prediction is a fundamental and challenging task for numerous applications, such as autonomous driving and intelligent robots. Current works typically treat pedestrian trajectories as a series of 2D point coordinates. However, in real scenarios, the trajectory often exhibits rand...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2021-12-01
|
| Series: | Computational Visual Media |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s41095-021-0236-6 |
| _version_ | 1818383915481563136 |
|---|---|
| author | Pei Lv Hui Wei Tianxin Gu Yuzhen Zhang Xiaoheng Jiang Bing Zhou Mingliang Xu |
| author_facet | Pei Lv Hui Wei Tianxin Gu Yuzhen Zhang Xiaoheng Jiang Bing Zhou Mingliang Xu |
| author_sort | Pei Lv |
| collection | DOAJ |
| description | Abstract Trajectory prediction is a fundamental and challenging task for numerous applications, such as autonomous driving and intelligent robots. Current works typically treat pedestrian trajectories as a series of 2D point coordinates. However, in real scenarios, the trajectory often exhibits randomness, and has its own probability distribution. Inspired by this observation and other movement characteristics of pedestrians, we propose a simple and intuitive movement description called a trajectory distribution, which maps the coordinates of the pedestrian trajectory to a 2D Gaussian distribution in space. Based on this novel description, we develop a new trajectory prediction method, which we call the social probability method. The method combines trajectory distributions and powerful convolutional recurrent neural networks. Both the input and output of our method are trajectory distributions, which provide the recurrent neural network with sufficient spatial and random information about moving pedestrians. Furthermore, the social probability method extracts spatio-temporal features directly from the new movement description to generate robust and accurate predictions. Experiments on public benchmark datasets show the effectiveness of the proposed method. |
| first_indexed | 2024-12-14T03:13:57Z |
| format | Article |
| id | doaj.art-51e8d24d414b4edea17e332a17e987d5 |
| institution | Directory Open Access Journal |
| issn | 2096-0433 2096-0662 |
| language | English |
| last_indexed | 2024-12-14T03:13:57Z |
| publishDate | 2021-12-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Computational Visual Media |
| spelling | doaj.art-51e8d24d414b4edea17e332a17e987d52022-12-21T23:19:12ZengSpringerOpenComputational Visual Media2096-04332096-06622021-12-018221322410.1007/s41095-021-0236-6Trajectory distributions: A new description of movement for trajectory predictionPei Lv0Hui Wei1Tianxin Gu2Yuzhen Zhang3Xiaoheng Jiang4Bing Zhou5Mingliang Xu6School of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversitySchool of Information Engineering, Zhengzhou UniversityAbstract Trajectory prediction is a fundamental and challenging task for numerous applications, such as autonomous driving and intelligent robots. Current works typically treat pedestrian trajectories as a series of 2D point coordinates. However, in real scenarios, the trajectory often exhibits randomness, and has its own probability distribution. Inspired by this observation and other movement characteristics of pedestrians, we propose a simple and intuitive movement description called a trajectory distribution, which maps the coordinates of the pedestrian trajectory to a 2D Gaussian distribution in space. Based on this novel description, we develop a new trajectory prediction method, which we call the social probability method. The method combines trajectory distributions and powerful convolutional recurrent neural networks. Both the input and output of our method are trajectory distributions, which provide the recurrent neural network with sufficient spatial and random information about moving pedestrians. Furthermore, the social probability method extracts spatio-temporal features directly from the new movement description to generate robust and accurate predictions. Experiments on public benchmark datasets show the effectiveness of the proposed method.https://doi.org/10.1007/s41095-021-0236-6trajectory predictionconvolutional LSTMtrajectory distributionssocial probability method |
| spellingShingle | Pei Lv Hui Wei Tianxin Gu Yuzhen Zhang Xiaoheng Jiang Bing Zhou Mingliang Xu Trajectory distributions: A new description of movement for trajectory prediction Computational Visual Media trajectory prediction convolutional LSTM trajectory distributions social probability method |
| title | Trajectory distributions: A new description of movement for trajectory prediction |
| title_full | Trajectory distributions: A new description of movement for trajectory prediction |
| title_fullStr | Trajectory distributions: A new description of movement for trajectory prediction |
| title_full_unstemmed | Trajectory distributions: A new description of movement for trajectory prediction |
| title_short | Trajectory distributions: A new description of movement for trajectory prediction |
| title_sort | trajectory distributions a new description of movement for trajectory prediction |
| topic | trajectory prediction convolutional LSTM trajectory distributions social probability method |
| url | https://doi.org/10.1007/s41095-021-0236-6 |
| work_keys_str_mv | AT peilv trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT huiwei trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT tianxingu trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT yuzhenzhang trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT xiaohengjiang trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT bingzhou trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction AT mingliangxu trajectorydistributionsanewdescriptionofmovementfortrajectoryprediction |