力合资本、合肥创新投等机构投资，Lux 系列视触觉传感器已实现规模化交付。
作者丨董子博
雷峰网(公众号：雷峰网)·鲸犀独家获悉，触觉智能企业“灵感机器人”新近完成数千万元天使轮融资，投资方包括力合资本、合肥创新投等机构，治臻咨询担任财务顾问。雷峰网·鲸犀了解到，公司创始人、CEO 张雅展博士毕业于港科大机器人专业，曾是腾讯资深 AI 研究员，主导过大规模数据管线及 AI 模型研发，后于国家实验室担任研究员、博导，在触觉感知与端到端精细操作模型的领域有多年研究，并在算法与硬件系统上均有落地经验。相比于触觉感知赛道的其他公司，灵感机器人入局更早，也有一套自己独创的“稠密力场”技术路线——在解算力场的同时，从同一次成像中提取接触区域的几何形态，并形成包含受力、形变、滑移与姿态要素的原生多模态的触觉数据，相较传统“强调几何重建”的方案，在信息密度和维度上均有很大的提升。在张雅展看来，让触觉真正参与操作，重点就是理解“交互力”——灵感机器人希望通过“稠密力场”路线，提升机器人的触觉感知与力控能力，来整体性地提升任务的执行能力。除了独特的技术选型，传感器的小型化、感知密度与工程化，也是灵感机器人的重要攻关方向。2019 年，团队就研发出了厚度 5mm 的超薄视触觉传感器，力场追踪的分辨率达到了每平方 10 万点以上。如今，公司发布的 Lux 系列视触觉传感器已经在多个领域实现了规模化落地，其具备超高力场分辨率、原生多模态、多维力感知等特性，并且依靠更小的物距与体积，可以适配机械臂、夹爪、灵巧手等硬件。而在 Lux 之外，灵感机器人还正在电子皮肤、六维力与新型触觉产品上探索，并准备向数采系统、数据管线和触觉智能系统等业务进军。据悉，公司本轮融资将被用于建设产线与优化量产能力，并进一步进行触觉智能系统的研发。
雷峰网原创文章，未经授权禁止转载。详情见转载须知。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Institutional investments such as pooled capital, compost innovation investments, and the Lux series visual touch sensors have been delivered on a scale.
The author, Dong Zibo.
Ray Peaks (WeChat Official Account:Ray Peaks Network) Whale alone learned that the Sensible Smart Enterprise “Inspirator Robots” has recently completed tens of millions of dollars in Angel Wheel financing, invested in institutions such as pooling capital, compost innovation, etc., and advised finance consultants. Ray Peak Whale learned that the founder of the company, Dr. CEO Zhang Yag-soo, had graduated from the Hong Kong Big Robotic Profession and had been a senior AI researcher in Tencent, leading the development of large-scale data pipelines and AI models, and then was a researcher at the National Laboratory, a guide, a multi-year researcher in the field of sensory and end-to-end precision operating models, and had landed experience in algorithms and hardware systems. Even earlier than other companies on the tactile track, the inspiration robots are entering the game, and there's a technological route of their own. – At the same time, the geometry of the area of contact is extracted from the same imaging, as well as the formation of original multi-modular tactile data containing elements of stress, shape change, slide and attitude, with a significant increase in information density and dimensions compared to the traditional “emphasis on geometric reconstruction” programme. In Zhang Yak's view, the focus of real involvement is on understanding “interactivity” - inspirational robots want to improve robots' sensory and control capabilities through the “breath field” route in order to enhance mission performance as a whole. In addition to unique technology selections, sensor miniaturization, sensor density and engineering are important thrusts for inspirational robots. In 2019, the team developed a 5 mm thick ultra-mislight touch sensor, with resolution tracking the force field reaching more than 100,000 points per square. Today, the Lux series of visual sensors of the company publish have been scalded in a number of areas, with superhigh field resolution, primary multimodular, multi-dimensional sensory characteristics, and can adapt to hardware such as mechanical arms, claws and smart hands, depending on smaller distances and volumes. In addition to Lux, inspirational robots are exploring electronic skin, six-dimensional power and new tactile products, and are preparing to advance into operations such as digital mining systems, data pipelines and tactile intelligence systems. It is understood that the current round of corporate finance will be used to build production lines and optimize production capacity and further develop tactile intelligence systems.
It's an original Ray Peak article, which is forbidden to reproduce without authorization. For further information, see the information below.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.