AIcrowdTime SeriesDates unconfirmed
AIcrowdTrajnet++ (A Trajectory Forecasting Challenge)
Official pageOpen the competition on AIcrowd↗
- Effort
- 16–40 hours
- Difficulty
- Advanced
- Compute
- Free GPU OK
- Scoring
- Automatic
About the task
Full rules ↗The organizer's text, in Chinese; there is no English translation yet.
TrajNet++要求从9帧已观察轨迹预测后12帧行人位置。官方明确Round 3继续接受研究提交，当前榜主指标为FDE（米，越低越好），同时展示碰撞与Top3指标。官方提供训练/测试ZIP、ndjson格式、PyTorch基线和评测工具。2020年ECCV截止仅适用于历史研讨会；2030平台轮次时间不作为当前报名截止，未承诺现金奖励。
- Scoring
- FDE (final displacement error; lower is better)
Is it for you?
Prepare before joining
Beginner fit35/100
Learning value80/100
The editors' notes, in Chinese; there is no English translation yet.
Who it suits
希望研究人群轨迹预测、社会交互和碰撞约束的机器学习或机器人方向开发者。
What makes it hard
需要严格保持观察/预测时间分割、处理多行人交互，并同时权衡距离误差与碰撞率。
Getting started
Skills to bring
In Chinese; there is no English translation yet.
- Python / PyTorch
- 时序预测
- 二维轨迹数据与ndjson
- 离线验证与防止时间泄漏
We haven't pulled tutorials, data or baselines from the official page yet; the task and the data are there.
First competition? Read How to Choose Your First AI Competition