SDSC Research Team Won the Second Place Prize at National Competition
A research team comprising two co-leaders from SDSC - Dr. Xinyue HE (2023 SDSC Ph.D. graduate and postdoctoral fellow of HKIDS supervised by Professor Lishuai LI) and Mr. Junyan SU (a 4-year Ph.D. student supervised by Professor Minghua CHEN) teamed up the other three Ph.D. students, two from SDSC including Mr. Enming LIANG (supervised by Professor Minghua CHEN) and Mr. Chuankai XIANG (supervised by Professor Lishuai LI), and the other member from HKU Mr. Zongqi HU, entered a national competition – “The 1st Low-Altitude Economy Flight Management Challenge 首屆低空經濟智能飛行管理挑戰賽”. They successfully notched up the 2nd prize for their project entitled “LEAP: Low-altitude Express AI Pilot” and received a prize award of RMB$50,000.
The competition focuses on the core issues of drone operation in urban low-altitude environments: multi-drone planning and scheduling. It features two main streams: Performance Competition (性能賽) and Creativity Competition (創意賽). The competition drew more than 200 participants, with over 71 teams entering from 36 universities in the Greater China region and overseas. SDSC’s team out-bided the other competitors and ranked no. 2 in the Performance Competition category. The competition is closely related to data science, as it requires complex urban environment data processing, real-time demand data analysis decision support, efficient scheduling and planning algorithm development. For Performance Competition, the participating teams face delivery scenarios in a defined map environment, where order demands are generated on the map based on Meituan's routine operations. The competition provides a simulation environment that can simulate the synchronized operation of multiple drones in an urban environment, and contestants should use these drones to finish delivery tasks.
The competition requires efficient scheduling algorithms to assign thousands of delivery tasks to tens of drones and planning algorithms to generate trajectories for drones in complex urban environments to finish the assigned delivery tasks. Successful and on-time delivery earns scores, while situations like overtime, collisions, and crashes lead to penalties on scores. Thus, contestants should develop and submit planning algorithms to efficiently and safely complete delivery tasks while avoiding collisions, battery depletion, and other issues.
The School encourages our students to join data science competitions to sharpen their skills, grow with this fastest-growing disciplines, and conduct impactful research in multiple domains of applications and theoretical development.
