Gasgoo Munich- RoboSense secured an order from a leading European humanoid robotics firm for its Active Camera (AC) series. The products are scheduled for mass production and delivery by 2026.
Launched in January 2025, the AC series uses a multi-sensor fusion architecture and an AI-Ready ecosystem. It aims to function as a "robot eye" for mobility and fine manipulation. The platform includes the AC1 and AC2 products.
Image source: RoboSense
The AC2 debuted at the 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2025).
The AC2 integrates solid-state dToF LiDAR, a binocular RGB camera, and an IMU. It is described as the industry's first "super sensor system" to deliver hardware-level 3D spatial perception and 6-degree-of-freedom motion data. The device features an accuracy of ±5mm across its range and captures 1600x1200 high-resolution RGB images. This allows robots to reconstruct 3D details and detect small objects, addressing challenges in fine manipulation.
The AC2 also features a wide field of view (FOV). It delivers a consistent 120°x90° view in fusion perception or individual sensor modes. This includes dToF, binocular vision, or RGB, and is more than 70% wider than traditional 3D cameras.
To support the AC2, the AI-Ready ecosystem offers specialized open-source algorithms for pose estimation and human skeleton recognition. Designed for developers, it provides open-source resources including the ACStudio toolkit, Wiki documentation, and datasets. The ecosystem supports platforms like Rockchip RK3588, NVIDIA Orin, and Horizon Robotics chips.
Measuring 102mm x 32mm x 45mm, the compact AC2 can be mounted in tight spaces. These include a humanoid robot's face, a robotic dog's head, or a mechanical arm tip. This ensures safe and efficient operation.
At CES 2026, RoboSense showcased a delivery robot equipped with the Active Camera series. The AC2 was mounted above the eyes and on the hand to provide close-range and high-precision stereo and color data. The AC1 was positioned on the body and back to handle long-distance navigation and obstacle avoidance.
The company pairs the system with its proprietary high-degree-of-freedom dexterous hand. It leverages the RoboSense AI platform for end-to-end operation and mobility. The platform covers core components, model training, data acquisition, and data flywheels. This enables fully autonomous operation and flexible, long-range manipulation.
