Recently，Easy Control Intelligent Driving Technology Team is well-known in the world.Argoverse 2 Sports Prediction Challenge，With the self-developed joint trajectory prediction framework，Won the first place in the world in the motion trajectory prediction list。

Argoverse 2 Sports Prediction Challenge is based onArgoverse 2 Motion Forecasting Dataset，It has been widely regarded as one of the important benchmark data sets for motion prediction in the field of automatic driving.。The event is an important event in the field of global automatic driving.，Attracted many top enterprises and research institutions to participate.，Jointly address technical challenges in autonomous driving。ArgoverseData set evaluation includes 7 technical indicators.，In this challenge，Easy to control the key indicators of the intelligent driving team——avgBrierMinFDEExcellent performance。

Argoverse2 Trajectory Prediction Challenge Ranking

Motion trajectory prediction plays a central role in automatic driving technology.，It's like a self-driving vehicle.“Eyes of the future”，By accurately predicting the surrounding vehicles、Dynamic changes of pedestrians and obstacles，Provide vehicles with timely、Accurate decision basis，Ensure driving safety，Optimize the driving path，Improve traffic efficiency，It is the intellectualization of automatic driving technology.、Security、The key to efficient driving。
In recent years，Many trajectory prediction models have been proposed.，However, there are still many difficulties that have not been fully solved in the existing schemes.。For example，The accuracy of the current method is low when dealing with interactive scenes.，The main reason is that joint prediction needs to consider the predicted trajectories of all targets at the same time.，This increases the complexity of the solution space.。In addition，The result of joint prediction depends on the stable and reliable prediction output of all targets.，To ensure the stability of the final result.。
In addition，Different from urban roads，Mine scenarios lack clear traffic rules and lane line constraints，Resulting in freer vehicle driving behavior。In addition，Mine vehicles are usually more than 5 meters wide.，Occupies a large space，Therefore, accurate behavior trajectory prediction is a necessary condition to ensure safe production.。This task of trajectory prediction，Pecially in open areas where multiple vehicles are operating，It puts forward higher requirements for prediction accuracy.。
To this，Combination of independent research and development of Easy Control Intelligent Driving TeamTransformerStructure of feature encoding and trajectory decoding for multi-vehicle interaction scene based on MATLAB，It can effectively solve the pain point of the existing trajectory prediction technology.。In addition，Easy control intelligent driving innovatively introduces the credibility supervision mechanism in the trajectory prediction algorithm.，And cyclic decoding combined with trusted modality selection，The accuracy of that prediction can be bet improved，And targeted to solve the needs of the mine scene。
 

Future，Easy control intelligent driving will apply this innovative technology to the practice of mine automatic driving.，In order to further improve the operation safety and efficiency of mine automatic driving.，Helping traditional mining industry to become intelligent、Green、Efficient、Transformation and upgrading of safe modern mining industry。