Songhao Piao

Harbin Institute of Technology

Papers

5

Total Citations

23

H-Index

4

About

Songhao Piao is a robotics and artificial intelligence researcher whose work spans multi-agent systems, reinforcement learning, robot path planning, and autonomous navigation. Over more than two decades of active research, Piao has consistently addressed fundamental challenges in making robotic systems more intelligent, adaptive, and cooperative. Among his early contributions, Piao tackled the computational inefficiencies of conventional reinforcement learning in multi-agent environments, proposing novel methods to improve cooperative behavior acquisition while reducing learning costs — work that has accumulated citations reflecting its foundational relevance. His parallel investigations into path planning introduced a hybrid genetic annealing algorithm that outperformed standalone genetic and simulated annealing approaches in dynamic multi-robot environments. More recently, Piao has advanced the frontier of imitation learning, developing frameworks that enable robots to continuously acquire new manipulation tasks while reusing prior experience — a critical capability for real-world deployment. His 2025 work on heterogeneous relational deep reinforcement learning for crowd navigation further demonstrates his ongoing commitment to solving complex multi-robot interaction problems. Practical contributions, such as GPRS-based guard robot systems, highlight his ability to bridge theoretical innovation with real-world applications, making his research portfolio both broad and enduring in its impact.

Research Focus

Key Achievements

4
H-Index
5
Papers
23
Total Citations
5
Avg Citations/Paper
🏆 Most Cited Paper
Fast reinforcement learning approach to cooperative behavior acquisition in multi-agent system
7 citations · 2003
📈 Most Prolific Year: 2003 (2 Papers)
🤝 Key Collaborators: 9
🏛 Institutions: Harbin Institute of Technology

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago