AI method: RNN,LSTM... boost
2019, LSTM
https://kth.diva-portal.org/smash/get/diva2:1375828/FULLTEXT01.pdf
2021 ,DP
Full range adaptive cruise control_11311-822928_Alippi.pdf (polimi.it)
Full-range Adaptive Cruise Control Based on Supervised Adaptive Dynamic Programming
分段线性拟合,车道线?分割点如何设置,却决于传感器的测量特性,局限性。 40m Lidar 精度高,以上就不行了,都需要考虑。
【自动驾驶算法】分段最小二乘法拟合 - 知乎 (zhihu.com)
自车heading ,通过Camerad的 C1 ,两边两个C1,三角关系,或者自车偏航角。
5次多项式
(17) (PDF) Adaptive Lane Change and Lane Keeping for Safe and Comfortable Driving (researchgate.net)
基于Frenet坐标系采样的自动驾驶轨迹规划算法研究 - 豆丁网
局部路径轨迹,AI ,规则方法。
A Review of Motion Planning for Highway Autonomous Driving (archives-ouvertes.fr)
2011.04752.pdf (arxiv.org)
Trajectory Planning for Autonomous Vehicles Using Hierarchical Reinforcement Learning
2004.05224.pdf (arxiv.org)
Result-level Fusion
Multi-level Fusion (hybird) use signal 2D box to get the ROI proposal ,选择需要的点云(gating)
Feature-level Fusion ( pixel level)
底层融合: 摄像头的信息painting到 lidar点云中。 --zeeker
大陆的演示
hybird fusion : 用2D框,去确定另一个 sensor ROI
目标级融合
multi-view fusion: BEV, front view ....tbd.
to read:
66] S. Vora, A. H. Lang, B. Helou, and O. Beijbom, “Pointpainting: Sequential fusion for 3d object detection,” 2019.
Song-Jingyu/PointPainting: This repository is an open-source PointPainting package which is easy to understand, deploy and run! (github.com)
[67] X. Chen, H. Ma, J. Wan, B. Li, and T. Xia, “Multi-view 3d object detection network for autonomous driving,” 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jul 2017. [Online]. Available: http://dx.doi.org/10.1109/CVPR.2017.691
自动驾驶|3D目标检测:MV3D-Net(一) - 知乎 (zhihu.com)
Camera OD:
ICCV 2021 Open Access Repository (thecvf.com) (calmcar)