STMicroelectronics integrates AI to MEMS sensors
Signal processing and AI algorithms have been combined in MEMS (micro electromechanical systems) sensors by STMicroelectronics. The Intelligent Sensor Processing Unit (ISPU) injects local decision-making while substantially saving space and power, says the company.
The ISPU combines a digital signal processor (DSP) able to run AI algorithms and a MEMS sensor on the same silicon. In addition to a reduction in size, compared to system-in-package devices, the ISPU is also claimed to cut power by up to 80 per cent. Merging sensor and AI puts electronic decision-making at the edge, said ST, where products enabled by smart sensors are able to sense, process, and take actions, in what has been called the Onlife Era, fusing technology and the physical world.
The Onlife Era acknowledges living with continuous assistance from connected technologies, enjoying natural, transparent interactions, and seamless transitions, with no discernible distinction between online and offline, ST explained. The ISPU allows the migration of intelligent processing into sensors that support the fabric of life, or as ST puts it: no longer at the edge but in the edge.
The proprietary low power DSP can be programmed in C, a language familiar to many engineers. It also allows quantised AI sensors to support full- to single-bit-precision neural networks. This ensures superior accuracy and efficiency in tasks such as activity recognition and anomaly detection by analysing inertial data, said ST.
“While technically challenging, integrating ST’s sensors on the same piece of silicon with our ISPU does improve sensor-based systems from an online experience to an Onlife one,” said Andrea Onetti, executive vice president, MEMS Sub-Group, at STMicroelectronics.
“It advances the sensor’s features to speed decision-making by reducing data transfers, enhancing privacy by keeping data local, while reducing size and power consumption, which cuts costs,” he added.
“Moreover, the ISPU is easily programmable with commercial AI models and can ultimately operate with all of the leading AI tools.”
ST’s proprietary, C-language-programmable DSP is an enhanced 32-bit RISC (reduced instruction set computing) machine. It is extensible (in the chip-design phase) for dedicated instructions and hardware components. The processor offers a full precision floating point unit, uses a fast four-stage pipeline, operates from 16-bit variable-length instructions, and includes a single-cycle 16-bit multiplier. Interrupt response is four cycles.
ST’s sensors with ISPUs will be packaged in standard 3.0 x 2.5 x 0.83mm packages and will be pin compatible with earlier models available from the company, for ease of upgrades.
ST also claims that combining the sensor and ISPU save five to six time power saving compared with system-in-package approaches in sensor-fusion applications. They also show a two to three times saving in Run mode.