Robotics Instruction
探索如何让机器人更好理解任务、约束和人类意图，让 instruction 成为可靠行动的起点。
ABOUT
HJH
I'm Jianheng Hu (胡健恒), a freshman at Beihang University.
I grew up in Chengdu and came to Beijing for university. The transition wasn't exactly what I had planned - but it's where I am, and I'm making the most of it.
My research interest is embodied intelligence: getting robots to understand instructions, generalize across environments, and act reliably in the real world. Right now I'm focused on sim-to-real transfer - the gap between what a robot learns in simulation and what actually works when you deploy it.
On the side, I've been working on a paper about Cloud-Edge-Device collaborative scheduling, the CCHIHH algorithm, which sits at the intersection of evolutionary computation and multi-armed bandits.
I also write small tools when I find something annoying enough to fix.
What I Explore
当前公开项目仍在打地基，研究方向会逐步聚焦到机器人指令、真实世界泛化和下一代机器人范式。
Sim-to-real Generalization
关注从仿真到真实机器人的迁移，让策略、数据和评估更能跨环境稳定工作。
New Robotics Paradigm
探索一种能把感知、控制、记忆、世界模型和泛化能力连起来的机器人范式。
How I Work
我会从小工具和可复现实验开始，把问题拆清楚，再把它们逐步连接到更大的机器人智能目标。
Keep In Touch
If you're working on robotics, sim-to-real, or embodied AI - or just want to talk - reach out.