Topics in Designing An Android Sensor Subsystem: Pitfalls and Considerations
This lecture will arm Android device architects with the tactical knowledge they need to navigate the Android Sensor subsystem and make knowledgeable design choices to improve user experience and improve battery performance. It will cover: Hardware architecture and trade-offs including latency, power, and software architecture implications: (Sensor selection, What is sensor fusion? Sampling Rates: Is more always better?), Wake up events and power considerations (Is a coprocessor the answer?), Gesture Detection Algorithm processing location and considerations (Where to perform gesture detection, Buy versus make, Porting algorithms to another processor, Testing human interface sensors), Integration into the Android Sensor Frameworks (Must have software checkpoints), Testing methodologies (Creating tools to aid develop and collect data). This talk targets the kernel/firmware developer responsible for the sensor architecture. They should be familar with kernel drivers, embedded systems, hardware bring up, Android services, and the C language. Jen Costillo Jen Costillo has been working in consumer embedded systems for almost 15 years. She has worked on network equipment, video set top boxes, biometric readers, as well as several human interface devices ranging from air mice to multi-touch devices to human gesture musical instruments. Her on-going projects include human gesture recognition to generate music in real-time and electronics wearables. She has given talks to San Jose State University ACM as well as She’s Geeky conferences 2007-2012. She currently works for Lab 126.