I-Corps: Semio: Customer discovery for a software platform for human-robotic interactions

这个i-Corps团队正在开发一个标准化的软件平台，允许人类通过语音和肢体语言等自然交流与机器人互动和使用。拟议的“Semio”软件平台允许机器人开发人员创建(通过开发人员工具)和部署(到应用商店)将由机器人执行的基于语音/手势的应用程序，并允许非专业机器人用户访问(从应用程序商店)和使用(通过自然通信和智能设备)这些机器人应用程序。拟议中的软件使机器人能够自动识别人类的语音和肢体语言，做出如何回应的决定(由应用程序的开发者指定)，并根据这些决定采取行动，同时还提供自己的语音和肢体语言。这种自然的交流界面增加了新手用户对机器人的易用性。Semio软件平台涉及新颖的算法和健壮的代码库，使机器人硬件平台能够基于一组核心社交行为来进行“开箱即用”的操作，这些行为作为面对面对话交互的基础；这些行为包括说话(所说的话)、眼睛凝视(注意力集中的地方)、手势(面部、头部、身体、手臂和手的动作)、话轮转换(何时说话和何时倾听)和人际间隔(在哪里进行对话)。拟议中的基于语音/手势的软件的目标市场是个人机器人，尽管该团队还在探索其他市场，如智能设备(智能手机、GPS和物联网)、游戏、交通(汽车和自动驾驶汽车)、传感器(微软Kinect和英特尔RealSense)以及通用机器人(工业、军事和商业机器人)。该团队目前的目标客户是这些互动技术的硬件和软件开发商。开发人员既可以是专家，也可以是非专家，他们可能是工业界、制造者社区、业余爱好者或学术界的成员。通过反复与客户面谈，该团队将继续将客户反馈融入到演示应用中，更好地改进产品内容和性能，满足客户识别的需求。该软件平台还将连接到硬件机器人平台，以开始向机器人市场的潜在客户进行概念验证演示并建立联系。

This I-Corps team is developing a standardized software platform to allow humans to interact with and use robots through natural communication, such as speech and body language. The proposed "Semio" software platform allows robot developers to create (through developer tools) and deploy (to an app store) speech/gesture-based applications to be executed by robots, and allows non-expert robot users to access (from the app store) and use (through natural communication and smart devices) those robot applications. The proposed software enables a robot to automatically recognize human speech and body language, make decisions about how to respond (specified by the developer for the application), and act based on those decisions while also providing its own speech and body language. This natural communication interface increases the usability of robots by novice users. The Semio software platform involves novel algorithms and robust code bases that enable a robotic hardware platform to operate "out of the box" based on a set of core social behaviors that serve as building blocks of face-to-face conversational interaction; these behaviors include speech (what is said), eye gaze (where attention is focused), gestures (face, head, body, arm, and hand movements), turn-taking (when to talk and when to listen), and interpersonal spacing (where to stand for to have a conversation). The target market for the proposed speech/gesture-based software is personal robotics, though the team is also exploring other markets, such as smart devices (smart phones, GPS, and the Internet of Things), gaming, transportation (cars and self-driving cars), sensors (Microsoft Kinect and Intel RealSense), and general robotics (industrial, military, and commercial robotics). The team is currently targeting customers who are hardware and software developers of these interactive technologies. Developers may be both expert and non-expert, and may be members of industry, the Maker community, hobbyists, or academia. Through repeated customer interviews, this team will continue to integrate customer feedback into the demo app to better improve the product content and performance to meet the identified needs of customers. The software platform will also be connected to a hardware robot platform to begin proof-of-concept demonstrations to and connection-building with prospective customers in the robotics market.