Infinin, 60GHz Radio Sensor Module Demonstration: Acquisition of radio wave authentication in the world's smallest class

ターンキーソフトウェアの活用で、システム開発を容易に

　Infinice Technologies Japan (hereinafter referred to as Infinin) is an ET & IoT 2021 (November 17-19, 2021, Pacifico Yokohama), the 60GHz Smart Radio Sensor Module "NJR4652" series and the "PSOC" MCU edge AI.We introduced demonstrations about "Modustoolbox Machine Learning (ML)" solutions to realize.The NJR4652 series was selected as the "IoT Technology Excellence Award" in the "ET & IoT AWARD 2021" by the organizer.

　The NJR4652 series has a built-in 60GHz radar IC made by an infinnion and an antenna, and a MCU 6, ARM Cortex-M4F based, which provides RF control and signal processing.The New Japan Radio has accumulated one package, including the peripheral parts.

　The external dimension of the package is 10.0 × 13.4x1.It is extremely small with 2mm.It is said to be the world's largest module product that has acquired the technical standard conformity that can be used in Japan.The 60GHz radar IC itself has already been adopted in smartphones.

　The NJR4652 series also has a turn -key software such as "Smart Entrance Counter Solution", which was developed by the Infinion, and "Presence Detection" for existence detection.For this reason, it is possible to quickly and easily develop an application system such as detection of entry and exit frequency at the entrance and detection of people in conference rooms.We will continue to develop and provide turn key software for various purposes, including New Japan Radio.

　At the booth, a demonstration was performed to measure the distance to the detection target using the NJR4652 series.The demo machine 60GHz radar IC has one transmission antenna and three receiving antennas.The demonstration machine has a frequency bandwidth width of 500MHz, and the distance resolution is several tens of centimeters."If you change the frequency bandwidth to be used with software, you can further enhance the distinction."

　In addition, we introduced the development environment Modustoolbox ML that realizes edge AI using PSOC MCU.The Modustoolbox ML offers middleware, software library, and ML exclusive tools for designers to evaluate and commercialize deep learning -based ML models.

　By using Modustoolbox ML, you can directly implement deep learning frameworks such as "TENSORFLOW" in PSOC.Moreover, it is possible to optimize the model for built -in platforms to reduce the size and complexity and verify the performance for test data.

　At the booth, we used a module that combines "PSOC PIONEER KIT" with a motion sensor to detect gestures.Specifically, we implemented learning data to detect the movement of "circular", "square", and "left and right", and showed the movement of the module in real time.

　In addition, we used the "PSOC 64 Standard Secure" MCU, which acquired the Level 2 authentication of "ARM PSA (Platform Security Architecture)", to enable safe updates to the firmware.PSOC 64 Standard Secure is suitable for applications such as personal healthcare devices and smart home security.The booth introduced a demonstration that correctly recognizes the input numbers while rewriting two types of data learned to be able to recognize the specific numbers.