Batty: AI-Driven Socialization Device for the Visually ImpairedBatty:面向视障人士的 AI 社交设备
Designer设计师
School project学校项目
4 weeks4 周
01 — The Opportunity & Challenge01 — 机遇与挑战
Defining Emotional Inequity in a Visual-Centric World在以视觉为中心的世界中定义情感不平等
55% invisible. 300 million people left out.55% 不可见。3亿人被排除在外。
In human communication, 55% of information is conveyed through non-verbal cues — facial expressions, micro-movements, and silent shifts in mood. For the 300 million people worldwide with moderate to severe visual impairment, this 55% is a "black hole" in social interaction.在人类沟通中,55% 的信息通过非语言线索传递——面部表情、微小动作与无声的情绪变化。对于全球 3 亿名中重度视障人士而言,这 55% 是社交互动中的"黑洞"。
While current assistive technologies focus on "functional navigation" (object detection, OCR), the emotional dimension of socialization remains largely ignored. Social interactions are often functional but shallow — leading to isolation and inequality.现有辅助技术大多聚焦于"功能性导航"(物体识别、OCR),却对社交的情感维度视而不见。视障人士的社交往往停留在功能层面,缺乏深度,由此带来孤立与不平等。
The Mission: Transform "invisible" emotions into "tangible" feedback — a real-time bridge for deep, empathetic connection.使命:将"不可见"的情绪转化为"可感知"的反馈——构建深度共情连接的实时桥梁。
02 — The Solution02 — 解决方案
A New Haptic Language一种新的触觉语言
Designing for multimodal interaction面向多模态交互的设计
The 5-Point Magnetic Matrix —五点磁性矩阵 ——A simplified 5-dot matrix mapping facial rotations and primary emotions into specific tactile patterns.简化的五点矩阵,将面部旋转与基本情绪映射为特定的触觉模式。
Form Follows Empathy —形式追随共情 ——Batty magnetically attaches to the back of a smartphone, using the phone's existing processing power while remaining socially unobtrusive.Batty 通过磁吸附在智能手机背面,利用手机现有算力,同时保持低调不突兀。
Ergonomic Grip —人体工学握持 ——Iteratively refined so fingers naturally rest on the haptic feedback zone during a video call.经过反复迭代,确保视频通话时手指自然落在触觉反馈区域。
03 — Technical Architecture03 — 技术架构
From Vision Algorithms to Physical Sensation从视觉算法到身体感知
A functional MVP validated with real users经真实用户验证的功能性原型
AI Vision Processing —AI 视觉处理 ——Facial landmark detection identifies emotional states and head orientations in real time.面部关键点检测实时识别情绪状态与头部朝向。
Low-Latency Transmission —低延迟传输 ——Data sent via Bluetooth to an Arduino Nano-controlled circuit.数据通过蓝牙发送至 Arduino Nano 控制的电路。
User Validation: Users correctly identified head rotations with 90%+ accuracy after minimal training.用户验证:经过少量训练后,用户识别头部旋转的准确率超过 90%。
04 — Strategic Vision04 — 战略愿景
Beyond the Device超越设备本身
A standard for inclusive socialization包容性社交的新标准
Scalability —可扩展性 ——Integration into mainstream platforms (TikTok, WeChat, Zoom) embedding AI-generated haptic metadata in every video stream.可集成至主流平台(抖音、微信、Zoom),在每条视频流中嵌入 AI 生成的触觉元数据。
Empathy-Driven Innovation —共情驱动的创新 ——Using AI & HCI to solve "soft" human problems — a shift from "assisting life" to "enriching life."运用 AI 与人机交互技术解决"软性"人类问题——从"辅助生活"转向"丰富生活"。
Video Showcase视频展示
The product form shown is a low-fidelity prototype. The final product form is feasible if technical scale challenges are overcome.视频中展示的为低保真原型。若能克服技术规模化挑战,最终产品形态具备可行性。