At last week’s Embedded World in Germany, Texas Instruments (TI) introduced Arm Cortex-based vision processors to add vision and artificial intelligence (AI) in applications such as video doorbells, machine vision and autonomous mobile robots.

The AM62A, AM68A and AM69A processors are supported by open-source evaluation and model development tools, and common software that is programmable through standard APIs, frameworks and models. This platform of vision processors, software and tools helps designers develop and scale edge AI designs across multiple systems while accelerating time to market.

“You need Intelligence at the edge,” said Artem Aginskiy, TI’s general manager for Arm processors. “A lot of sensors are around vision.”

These processors feature a system-on-a-chip (SoC) architecture that includes integrated components such as Cortex-A53 or Cortex-A72 central processing units, a third- generation TI image signal processor, internal memory, interfaces, and hardware accelerators that deliver from 1 to 32Tops of AI processing for deep-learning algorithms.

The AM62A3, AM62A3-Q1, AM62A7 and AM62A7-Q1 support one to two cameras at less than 2W in applications such as doorbell cameras and smart retail systems.

The AM68A enables one to eight cameras in applications such as machine vision, with up to 8Tops of AI processing for video analytics. And the AM69A achieves 32Tops of AI processing for one to 12 cameras in applications such as edge AI boxes, autonomous mobile robots and traffic monitoring systems.

Beginning in Q2 2023, designers can access a public beta of TI’s free open-source tool Edge AI Studio. This web-based tool allows users to develop and test AI models quickly using user-created models and TI’s optimised models, which can also be retrained with custom data.

TI also announced a 32bit general-purpose MCU portfolio. The scalable Cortex-M0+ microcontrollers have a wide range of computing, pinout, memory and integrated analogue options.

“TI is building the industry’s most comprehensive portfolio of Arm Cortex-M0+ based MCUs, expanding an already extensive semiconductor offering with options for general-purpose designs,” said Vinay Agarwal, TI vice president. “Our new MCUs provide the flexibility our customers need to enhance the sensing and control capabilities of their systems while cutting cost, complexity and design time.”

Designers can select computing options from 32 to 80MHz with maths acceleration and multiple configurations of integrated analogue signal-chain components, including a zero-drift operational amplifier on an MCU and precision 12bit, 4MS/s analogue-to-digital converters. This flexibility helps designers meet their current design requirements and plan for future designs, all within the same MCU portfolio.

“You want to reuse hardware and software in multiple generations,” said Agarwal. “We aim to simplify embedded system design.”

More than 100 MCUs are planned this year, with software, design support resources and coding tools – including graphical tools that streamline device configuration – created to help designers code once and then scale across future MSPM0-based designs.

A software development kit (SDK) provides a cohesive experience that includes a wide variety of drivers, libraries, more than 200 code examples and subsystem reference designs.

In addition to a software, tool and training ecosystem, all analogue and embedded processing parts are supported by TI’s internal manufacturing to help meet demand for decades to come.

At Embedded World, visitors to TI’s stand saw how MSPM0 MCUs could increase system efficiency and processing and sensing capabilities in applications such as chamberless smoke detectors, blood pressure monitors, pulse oximeters and motor-control systems.

These MCUs are available in multiple package sizes with 16- to 64-pin package options and flash memory ranging from 16kbyte to 128kbyte.