Hand-held camera cube reference design brings AI to the edge

Artificial intelligence (AI) which was previously limited to expensive machines with large power budgets can now be embedded in space-constrained, power-powered edge devices. Maxim Integrated says its MAXREFDES178# camera cube executes low latency AI vision and hearing inferences on a coin cell power budget with reduced cost and size.

The MAXREFDES178# enables low power IoT devices to implement hearing and vision. It is based on the MAX78000 low power microcontroller with neural network accelerator for audio and video inferences. The system also contains the MAX32666 low power Bluetooth microcontroller and two MAX9867 audio codecs. The system is delivered in a compact form factor to show how AI applications, such as facial identification and keyword recognition, can be embedded in low power, cost sensitive applications such as wearables and IoT devices.

AI applications require intensive computations, which is usually performed in the cloud or in expensive, power-hungry processors: self driving cars is an example. Maxim says that its MAXREFDES178# camera cube demonstrates how AI can operate on a low power budget, enabling applications that are time- and safety-critical to run on even the smallest of batteries. The MAX78000’s AI accelerator slashes the power of AI inferences up to 1,000x for vision and hearing applications, compared to other embedded solutions, reports Maxim. The AI inferences running on the MAXREFDES178# also show dramatic latency improvements, running more than 100x faster than on an embedded microcontroller.

The compact form factor of the camera cube (1.6 x 1.7 x 1.5inch of 41 x 44 x 39mm) allows AI to be implemented in wearables and other space-constrained IoT applications. The MAX78000 is up to 50 per cent smaller than the next-smallest GPU-based processor, says Maxim, and does not require other components like memories or complex power supplies to implement cost-effective AI inferences.

The MAXREFDES178# and the MAX78000 is available now, together with the MAX32666GWPBT+T RF microcontroller and the MAX9867EWV+T stereo codec   at Maxim Integrated’s website and authorised distributors.

http://www.maximintegrated.com

> Read More

Mouser Electronics signs global distribution agreement with Telink Semiconductor

Low power radio frequency and mixed signal chips for the IoT from Telink Semiconductor are now available from Mouser Electronics, following the signing of a global distribution agreement.

Mouser will add Telink’s range of SoCs, starter kits and multi-protocol IoT products to its portfolio. Among the highlights of the range are the Telink TLSR827x series of low power, concurrent, multi-protocol IoT devices operating at the ISM 2.4GHz band. The devices support Bluetooth Low Energy 5.1, Bluetooth SIG Mesh, Zigbee, HomeKit, RF4CE, and 2.4GHz proprietary protocols. They are built on a 32-bit RISC-based microcontroller capable of running up to 48MHz and feature up to 32 general-purpose inputs/outputs (GPIOs), including a 14-bit ADC, I²C, SPI, I²S, and stereo audio output.

There is also the TLSR8278 audio RCU (remote control unit) starter kit features a button function and voice command function support and is therefore suitable for RF and IR remote control applications. The starter kit has 27 valid buttons and two-colour indicating LEDs. It also offers Google voice service.

Mouser also offer the TLSR825x series multi-protocol single-chip devices suitable for IoT and human interface device (HID) applications. The low power, concurrent devices support Bluetooth Low Energy 5.0 and up to eight antennae for indoor positioning in addition to standard Bluetooth 5.0 features. Designers can use the TLSR825X devices in a wide range of applications, including wearable devices, wireless toys, advanced remote controls, smart lighting, and smart home products.

Telink’s Kite Mesh Starter Kit is based on the TLSR8258 and features multiple PCBAs to implement Bluetooth Mesh applications. It is suitable for home automation and smart lighting, and includes multiple USB dongles for emulating gateways or mesh nodes, as well as a remote control PCBA to emulate remote controllers or wall switches. The mesh starter kit is supported by a range of PC tools and a software development kit to support rapid development of mesh applications.

https://www.mouser.com

> Read More

Compact battery chargers are dedicated to charging in tight spaces

As wearable, medical and smart sensor products use streamlined power management for energy harvesting hardware, e-peas has developed a family of battery charger devices, dedicated to charging. The company’s existing power management ICs (PMICs) offer direct power delivery to the application, as well as the charging of energy storage elements (such as Li-Ion batteries or super capacitors), but the AEM10900, AEM10300 and AEM30300 are fully dedicated to the charging function. This allows them to be used where a simpler implementation can be used and where there are space or cost constraints.

The PMICs achieve zero quiescent current draw from the battery. In other words, if energy harvesting stops for a prolonged period of time, the energy stored into the battery will not be wasted supplying the PMIC.

Working in conjunction with a single-cell photovoltaic panel, the integrated AEM10900 PMIC boost converter is optimised for solar-based energy harvesting implementations. It has a fast maximum power point tracking (MPPT) functionality designed for objects in movement. This allows the device to get the most energy from the ambient illumination available, storing as much as possible. It also has a 250mV cold start capability, which means it can commence with charging the battery even when light intensities are very low. AEM10900 introduces an I2C interface to minimise the pin count and to offer a larger set of potential configurations. The PMIC includes battery thermal protection, a joule counter to let the user know the amount of energy harvested, and a shipping mode in which the battery cannot be charged. The AEM10900 is suitable for wearable consumer products and body-worn medical monitoring equipment.

The AEM10300 and AEM30300 PMICs both have built-in low power DC/DC converters supporting operation over an input voltage range of 100mV to 4.5V. Adaptive energy management permits these devices to automatically switch between boost, buck-boost and buck operational configurations as deemed appropriate. This ensures that optimal energy transfer is always maintained between the respective inputs and the storage element.

The AEM10900, AEM10300 and AEM30300 battery chargers PMICs only require three external components, advises e-peas, for energy harvesting to be added while keeping the bill of materials costs low and taking up very little board space.

Geoffroy Gosset, CEO and co-founder of e-peas said: “Following on from in-depth consultations with our customer base, it became clear that having compact solutions for charging only was going to be of real value.”

The AEM10300 and AEM30300 PMICs are supplied in a 28-pin QFN package format, measuring 4.0 x 4.0mm and the AEM10900 is available in either the 28-pin QFN package or a 16-pin WLCSP (with 2.0 x 2.0mm dimensions).

http://www.e-peas.com

> Read More

Payment bracelets interpret gestures and use biometric data

Collaboration between Italian start-up, DEED and Infineon Technologies, is showcased at MWC21. The get bracelets interpret human gestures and use biometic data to pick up a call or make payments.

At the core of get is a system consisting of components from Infineon that enable the wearable with connectivity, computing, sensing and security capabilities. Infineon’s Secora Connect supports the payment functionality based on lowest power consumption to achieve longest battery life for the consumer. Infineon’s Xensiv MEMS technology provides high-fidelity voice recording during phone call. The PSoC 6 microcontroller family which uses a high performance dual-core M4/M0 architecture is paired with Infineon’s Airoc Wi-Fi and Bluetooth for secure, low power  connectivity.

According to Edoardo Parini, CEO and founder of DEED, the bracelet’s pioneering features include new and higher security standards, an ID acquisition method for contactless payment and there is no screen. “It is the perfect bridge between ‘you‘ and ‘your’ digital-self!” said Parini.

Patented techniques have been used to create a seamless, light and water resistant wearable wristband, made up of several layers, based around a rigid-flex PCB. The intuitive human machine interface (HMI) allows for natural operation because the wearer does not have to swipe on screens or touch any display. Motion sensors with artificial intelligence (AI), for gesture recognition allow it to interpret human gestures, for example, to pick up a call, to check the time or to make payments. Consumers can use it to listen to audio or answer calls by holding their finger to their ear by ‘wrist bone conduction’, sending the sound through the body to the inner ear. Contactless payments can be released after individual electrocardiogram-based biometric identification. The bracelet also allows fitness and health monitoring.

http://www.infineon.com

> Read More

About Smart Cities

This news story is brought to you by smartcitieselectronics.com, the specialist site dedicated to delivering information about what’s new in the Smart City Electronics industry, with daily news updates, new products and industry news. To stay up-to-date, register to receive our weekly newsletters and keep yourself informed on the latest technology news and new products from around the globe. Simply click this link to register here: Smart Cities Registration