Hewlett-Packard has developed a sensor platform that taps its experience as the largest supplier of the tiny mechanical structures in inkjet printheads. HP claims its MEMS process technology enables sensors a thousand times more sensitive than the models in use today, which were designed to trigger airbags during the violent motion of an automobile accident.
HP is prototyping a multisensor module, using its accelerometer with other sensors, that it hopes to market for all types of wireless sensor networks.
“We want to create a multisensor-module chip that measures it all—vibration, sound, light, temperature, pressure, humidity and chemical signatures—[and] crank out enough of these to cover the world with about a trillion nodes,” said Peter Hartwell, a senior researcher at HP Labs.
HP's prototype of a sensor network node, using its own MEMS accelerometer, that will be deployed as part of HP Labs' Central Nervous System for Earth (CeNSE).
Photo: Margie Wylie
Cross-section of HP's MEMS accelerometer, which uses a three-wafer chip stack with a large proof mass in the middle to achieve its ultrasensitivity.
HP argues that even the most complex sensor can be cost-reduced by mass producing it on a scale rivaling that for the disposable electromechanical structures in inkjet printheads. HP’s current deal with Shell Oil to disperse ultrasensitive seismic sensors could drive such volume manufacturing, but consumer markets ultimately would be the best avenue to global domination.
“The more sensors you add, the more things that apps will be able to measure that you didn’t think of or expect.” said Hartwell.
Eventually, the data streaming in from everybody’s pocket sensors would be integrated with traditional sources. For instance, a building’s smart thermostat could access the mobile phone’s temperature sensors to find hot or cold spots within the building.
If the cost was truly low and the unit very small I could see the use of these devices in many places: every phone in a company (air quality,temp,humidity,fire,smoke, etc.), every room in the house (same list of monitored parameters), packaged with shipments of food/drugs to ensure safety/freshness (is already done to some extent), buried in fields/lawns to monitor water/soil conditions/bug infestations, just to name a few things. One thing that I did not see in the article was how to transfer or collect the data for use off sensor? This is needed to keep costs low, perhaps 2.4Ghz mesh networks?
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.