论文arXiv 2611.260742 天前
宇树H1-2验证生物脉冲网络步态控制e-prop达100%存活率
本文由 AI 根据公开内容整理,可能有误,请以原文为准:Controlling the Humanoid Robot H1-2 With a Simulated Biologically Plausible Network(arXiv 2611.26074)
核心内容
针对生物脉冲神经网络(SNN)在人形机器人运动控制中兼顾生物合理性与鲁棒性的难题,研究团队提出受生物启发的SNN架构并在宇树H1-2全尺寸人形机器人上进行实机验证。实验通过强化学习专家策略蒸馏与数据集聚合训练SNN,对比了固定储备池、奖励调制STDP及e-prop三种循环权重学习机制。结果显示,e-prop方案达到100%存活率,性能完全匹配非生物基准BPTT,而生物保真度更高的固定储备池和STDP方案存活率分别仅为43.7%和11%。
技术实现上,系统采用RL专家策略蒸馏结合DAgger数据集聚合方法训练SNN控制器。e-prop(资格传播)机制成功解决了SNN不可微导致的梯度估计难题,在保留脉冲时序依赖特性的同时实现了高效优化。尽管固定储备池模拟了体外神经培养物特性且STDP符合赫布学习规则,但在全尺寸双足 stepping 任务中遭遇性能瓶颈。该研究首次在物理实体上证明了特定类脑学习算法可替代传统反向传播,为具身智能的神经形态控制提供了关键实测数据。
关键看点
- e-prop性能对标:资格传播算法在H1-2步态控制中实现100%存活率,性能持平非生物基准BPTT,验证了类脑优化器在复杂本体上的有效性。
- 生物机制性能瓶颈:高生物保真度的固定储备池与奖励调制STDP方案存活率分别仅为43.7%和11%,揭示了纯生物启发机制在当前硬件下的局限。
- 训练范式创新:采用RL专家策略蒸馏结合数据集聚合(DAgger)方法训练SNN,解决了脉冲网络直接强化学习效率低的问题并支持真机部署。
- 全尺寸实机验证:突破SNN仅在仿真或小型平台验证的现状,首次在宇树H1-2全尺寸人形机器人上完成双足步进测试,提供关键物理实测数据。
工程与产业观察
该研究确立了e-prop作为连接计算神经科学与机器人控制的桥梁地位。对于具身算法工程师,表明在追求低功耗神经形态芯片部署时,纯生物机制尚需妥协,混合优化策略是当前最优解;产业界应关注SNN从仿真走向全尺寸本体的信号处理延迟与算力适配问题,这决定了类脑控制在人形机器人量产中的实际落地窗口期。
英文摘要 (Abstract)
Artificial Spiking Neural Networks have gained remarkable attention in the last few years due to their proven ability in simulating biological neural cultures and producing similar activity and trends. Nevertheless, whether they can be trained for humanoid locomotion while retaining biological plausibility and robust performance at once remains an open question. We study this problem by proposing a biologically inspired spiking network, trained by different learning schemes and tested on the H1-2 Unitree Humanoid. These schemes differ in how recurrent weights are trained. Ordered from most to least biologically faithful, they are: a fixed reservoir mimicking an in-vitro neural culture, where recurrent weights are kept frozen and only a linear readout is trained, the same network but with its synapses trained by reward-modulated spike timing dependent plasticity and eligibility propagation, and finally backpropagation through time as a non biological baseline for comparison. All schemes relied on distillation of a reinforcement learning expert policy with dataset aggregation. We show that e-prop reached 100% survival rate matching backpropagation through time, whereas both the fixed reservoir and the reward-modulated STDP, though the closest to biological cultures, plateaued at a much lower performance (~43.7% and ~11% survival). We explore in further detail which trends a real culture could and could not produce until now. Finally we demonstrate brief bipedal stepping on a full size humanoid H1-2.
作者:Fatima Atwi (LAAS-GEPETTO, Nantes Univ - ECN), Constant Roux (LAAS-GEPETTO), Christian Bergaud (LAAS-MEMS), Olivier Stasse (LAAS-GEPETTO)
外部来源
本页事实对应的公开页面,以原文为准。(点击将离开本站)