Sensor Expo 2017 will be in San Jose, CA for the 2nd year in a row! I am excited that the event organizers are staying at the hub of innovation for electronics. This will be my 5 year of host the IoT pre-conference program. The theme last year was “IoT 2.0 – sensor innovation moves from smart to intelligent”.
The key takeaways from last year’s speakers
Google (Steve Malkos, Technical Program Manager) – Android sensor framework support 32 different sensors, + they have the added the flexibility to add external sensors through dynamic sensor discovery
Yole Development (Guilliame Giradin, Market & Technology Analyst) – offered several insights
1) MEMS and Sensor transitioning to 3 main hubs: Inertial, Environmental and Optical.
2) warns that many of the MEMS industry is digging its own grave. He offered mainly 3 options to escape the madness.
• Innovate the sensor – increase features or capability, or reduce costs
• Integration with other MEMS devices - Combine with other sensors
• Software and Sensor Fusion
Rhiza (Josh Knauer, President/CEO) – shared a simple message – reaching the consumers is harder than ever. The retail industry will spend over 1.1 Trillion to determine: age, gender, income, education, and residence. Why not spend the technology? I’ve always said you need to follow the money to get the answers you seek.
Intelli-vision (Vaidi Nathan, President CEO) – was a timely follow. The answer to the retail industry is cameras! Yole talked about software and cameras and now Intelli-vision is highlight what software can do to enable cameras. Vision as a service is a great title as the use cases are so many when it comes to cameras. Clearly, smart retail is big use case, there is also smart cities, smart homes and smart machines. These intelligent cameras enable predictive and intuitive actions, contextual awareness and insight.
Panel discussion was on sensor innovation. We had some big name players like Knowles (Mike Adell, VP Marketing Intelligent Audio), ST Micro (Edoardo Gallizio, Sensor Product Marketing), and Bosch Sensortec (Marcellino Gemelli, Director of Global Biz Dev.) and to spice things up I seated them next to each other. We sprinkled in Rhiza and Intelli-vision to keep the panel from going into reality TV mode. You never know when you put these die-hard competitors on stage. The audience was receptive and several questions were fielded, but I have to say I was a little disappointed that there were no secrets revealed or even hinted at. Next year, I am going to negotiate that each panelist will have to share something insightful to get on the panel.
Quanergy (Louay Eldada, CEO/co-founder) – was in the pre-con, and certainly says cameras are great but so are Lidars. We can see that solid state version announced in 2016 will reach sub $250 but the next Gen solid state will get to sub $100! I expect we will be seeing a lot of these Lidars in autonomous drive vehicles. The best thing that was highlighted is that these Lidars will not compromise the look of a car, meaning no dome required on the top of those test cars we see today.
Gestigon (Moritz von Grotthuss, CEO) – Following the lead of Quanergy, Gestigon used automotive as a use case to showcased how gesture be used as a natural user interface. I have to say after seeing the presentation I have an appreciation on how you can capture motion through body modelling and pixel classifier. By the way, one interesting tidbit that was shared: Self-driving features could lead to more sex in moving cars (source Canadian Automated Vehicles Centre of Excellence)
NXP (Mike Stanley, Systems Engineer) offered a lesson on sensor data analytics. If you were thinking of using sensor data and create a machine learning model. Mike took you through the process step-by-step. Matlab’s Classification learner to IBM’s SPSS Model to Microsoft’s Azure Machine Learning tools
Pixart (Charles Chong, Strategic Technical Marketing) offered up lessons on CMOS Image Sensors can do way more than just take pictures. What was interesting to me in this session was the power vs resolution vs Frame Rate. Market driven by Always ON, Simultaneous Localization and Mapping (SLAM) and New forms of Human Machine Interface (HMI). This certainly fit what most of the earlier speaker from Yole, Intelli-vision, and Gestigon had to say.
The last speaker, AmbiqMicro, (Scott Hanson, CTO) was the perfect closer. If we are to make smart devices intelligent we will need ultra-low power sensor processing. Scott highlights that processing can happen at 3 different levels: Cloud, Hub, or device. His projects that when a device goes from Smart to intelligent it will likely need reduce battery life by 20x.
However he offers a combination of sub-threshold IC design techniques coupled with proper software techniques will help overcome 20x decrease. Is final words were:
• IoT 0.1 requires ultra-low power sensors
• IoT 1.0 requires ultra-low power processing
• IoT 2.0 needs all of the above with intelligent system design
Those were the takeaways from 2016. Stay tuned for the agenda for 2017. I am looking to change things up. Some of the ideas I am thinking of are: a panel that will share something “new”, product giveaways like development boards in 2014, and prizes drawing like the drone in 2015.
Discount code for registering for Sensors Expo 2017 is: DSP100.