Apple Research Unveils HALO Autonomous Control System for Humanoid Robots
苹果，也盯上了机器人“大脑”
Apple Research has unveiled HALO, an autonomous control system for humanoid robots, demonstrated on Unitree G1. The system integrates perception, task planning, and motion control, utilizing a Mac mini M4 Pro and the Qwen3.6-35B-A3B model for local execution. Trained via simulation in 1,000 virtual environments to reduce reliance on teleoperation, HALO achieved an average success rate of about 91% across 360 single-task tests in six unscanned real-world settings. While complex chained tasks showed a 70% success rate in one environment, the team has open-sourced code for data generation and deployment.
Source: Chinese robotics — Research news discovery · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Control
- Reported autonomousOpen source S5
Reported numbers
training_hours
≈300 hours
Reported success rate
≈91%
Reported trials
Reported success rate
70%
Reported trials
task_attempts
360
Reported trials
virtual_environments
1,000
- code: not_reported
- paper: not_reported
- dataset: not_reported
- weights: not_reported
What remains unknown
Not established in the collected evidence: Environment, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
近日，苹果研究团队公开名为 HALO 的人形机器人自主控制系统，并将其部署在宇树G1上验证。
Open source S5
HALO尝试让机器人凭借自己的摄像头画面完成这些判断，并在执行过程中持续修正动作。
Open source S9
研究人员生成了1000个虚拟家庭环境，制作约300小时的主要训练数据，再将学到的能力带到真实的宇树G1上。
Open source S18
苹果团队给出的测试结果是，在6个未提前扫描或建图的真实环境中，进行360次单项任务测试，平均成功率约91%。
Open source S10
论文显示，将几个动作串起来执行时，真实环境中的平均成功率为70%，这部分测试仅在一个真实环境中开展。
Open source S22
这次，替它走进真实房间、坐椅子、推箱子的，是宇树G1。
Open source S4
这台背着Mac mini的宇树G1，让苹果的机器人“大脑”研究有了一个看得见的载体。
Open source S24
Source：Chinese robotics — Research news discovery · news.qq.com