MFTraj: Map-Free, Behavior-Driven Trajectory Prediction for Autonomous Driving

Authors: Haicheng Liao, Zhenning Li, Chengyue Wang, Huanming Shen, Dongping Liao, Bonan Wang, Guofa Li, Chengzhong Xu

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Evaluations on the Argoverse, NGSIM, High D, and Mo CAD datasets underscore MFTraj s robustness and adaptability, outperforming numerous benchmarks even in data-challenged scenarios without the need for additional information such as HD maps or vectorized maps.
Researcher Affiliation Academia Haicheng Liao1 , Zhenning Li1 , Chengyue Wang1 , Huanming Shen2 , Dongping Liao1 , Bonan Wang1 , Guofa Li3 , Chengzhong Xu1 1University of Macau 2University of Electronic Science and Technology of China 3Chongqing University
Pseudocode No The paper describes its methods using prose, equations, and architectural diagrams, but it does not include pseudocode or a clearly labeled algorithm block.
Open Source Code No The paper does not provide any explicit statement about making its source code publicly available or include a link to a code repository.
Open Datasets Yes Datasets. We tested model s efficacy on Argoverse [Chang et al., 2019], NGSIM [Deo and Trivedi, 2018], High D [Krajewski et al., 2018], and Mo CAD [Liao et al., 2024b] datasets.
Dataset Splits No The paper describes data segmentation for observation and prediction horizons but does not specify explicit training, validation, or test dataset splits (e.g., percentages or counts).
Hardware Specification Yes We implemented our model using Py Torch and Py Torch-lightning on an NVIDIA DGX-2 with eight V100 GPUs.
Software Dependencies No The paper mentions 'Py Torch and Py Torch-lightning' but does not specify version numbers for these software components.
Experiment Setup Yes Using the smooth L1 loss as our loss function, the model was trained with the Adam optimizer, a batch size of 32, and learning rates of 10 3 and 10 4.