DAT: Deep Learning-Based Acceleration-Aware Trajectory Forecasting
Ali Asghar Sharifi, Ali Zoljodi, Masoud Daneshtalab, Silvia Liberata Ullo, Li Zhang · 2024
Read DAT: Deep Learning-Based Acceleration-Aware Trajectory Forecasting by Ali Asghar Sharifi, Ali Zoljodi, Masoud Daneshtalab, Silvia Liberata Ullo, Li Zhang (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11677107.txt.
Description
Sharifi and co-authors present in 2024 a deep learning model that spots road users and predicts where they will move. Safe self driving needs such forecasts to avoid crashes. Their model works straight from raw sensor data and gives a forecast at each time step.
The key new part uses speed change, not just speed and place, to guide each forecast. The team also built a way to work out true speed change for training, and a spotter that reports it for each object. On a large shared driving dataset, the model beat past methods, with the best gains for objects that speed up, slow down, or turn in curved paths.
Catalog record
| CREATOR | Ali Asghar Sharifi, Ali Zoljodi, Masoud Daneshtalab, Silvia Liberata Ullo, Li Zhang |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |