Shenzhen-based Manifold Tech raised a seven-figure RMB pre-Series A round, backed by Hony Capital.The funds will be used to support custom production of core components, scale up product deployment, and broaden market reach. Earlier, the company raised seed funding from ZhenFund and received angel investment from Junsan Capital.
Deep Tech Roots: Manifold Emerges from University Lab to Advance 3D Perception for Robots and Drones
Founded in 2022, Manifold specializes in developing 3D perception and reconstruction algorithms for robots and drones. These systems let machines perceive and recall their surroundings for real-time interaction. Manifold’s founders come from HKU’s MaRS Lab, experts in drone navigation and LiDAR SLAM. Lab director Professor Zhang Fu, a former DJI advisor and Livox co-founder, now guides Manifold’s technical strategy.
With advancements in AI and machine learning, 3D sensing systems have become faster and more precise.High-performance platforms now process large image and point cloud data efficiently, enabling real-time environment reconstruction. These technologies are rapidly gaining traction across multiple industries.
High-resolution 3D models support cultural preservation through virtual exhibits and education, while in emergencies, 3D data helps map fire scenes, trace ignition points, and assess structural damage for faster, informed responses.
Navigational Challenges Facing Autonomous Robots Today
Spatial awareness and memory are vital for robotics, enabling safe, efficient navigation in complex environments. Yet many current systems have significant limitations. GPS is unreliable indoors, and technologies like UWB and Bluetooth depend on fixed infrastructure. Meanwhile, visual and LiDAR-based navigation can struggle in unfamiliar or constantly changing surroundings, often resulting in navigation errors.
Hardware limitations also pose problems. Manual calibration is labor-intensive, and any change to the setup or environment usually demands expert-level parameter adjustments. Teams must also manage large volumes of data and continually optimize algorithms for different applications, which adds cost and delays.
To overcome these challenges, Manifold developed MindSLAM—a solid-state, multi-sensor fusion system built on proprietary algorithms. At its core is Odin 1, the first module to combine spatial perception and memory for advanced robotic navigation.
Odin 1 features a SPAD dTOF depth sensor, high-res color camera, and IMU. It synchronizes spatial and temporal data across all sensors to produce highly accurate and stable results. Could you clarify what you mean by “sh“? Are you looking to shorten something or referring to something specific? It also generates 700,000 point cloud data points per second, significantly enhancing the detail and completeness of spatial data.
Odin 1 as a Robotic Hippocampus
Functionally, Odin 1 mimics the hippocampus in biological organisms—responsible for processing spatial memory. Odin 1 merges spatial and temporal data to create detailed 3D maps, allowing real-time object detection and mapping even in low-light or sparse environments.
These features boost autonomous navigation, enable more effective route planning, and improve robotic performance in complex or dynamic environments.
In addition, MindSLAM is linked with Manifold’s MindCloud platform, which enables users to instantly convert real-world spaces into photorealistic 3D simulations. The platform supports the creation of digital twins, manages spatial data, and facilitates simulation-based training for robotics algorithms—forming a comprehensive foundation for both development and operational planning.
According to 36Kr, Manifold’s real-time, true-color 3D reconstruction technology is already being applied across industries such as construction digitization, renovation surveying, fire scene modeling, traffic accident analysis, and industrial manufacturing.
In real-world scenarios, drones and robots using Odin 1 can enter disaster areas and produce live 3D maps, giving emergency teams detailed structural insights. On construction sites, the device can monitor spatial developments, track project progress, and evaluate construction quality—streamlining workflows and reducing manual labor.
Manifold is also working with multiple robotics companies to make intelligent sensing modules more affordable and provide end-to-end solutions for navigation, mapping, and localization. Mass production of Odin 1 is planned for July, followed by a global launch.
Read the original article on: Kr-Asia
Read more:Sources say LangChain is on the Verge of Achieving Unicorn Status