AI at the Edge: Smarter, Faster, and More Secure with Gesture and Voice Control
Aaron Blutstein 6 March 2026 
With expertise, foresight, and a deep understanding of the demands of emerging markets, the specialists at Rutronik System Solutions are advancing the development of existing base and adapter boards. Their work extends beyond refinement to seamless integration into application-driven demonstrations. A prime example of this innovation is an AI control system featuring gesture recognition via the RAB3 radar and keyword detection powered by Infineon’s PSOCTM Edge E84. This microcontroller enables decentralised neural network processing “on the edge,” enhancing the system’s speed, stability, and security.
The Intelligence Behind the Demo
At the heart of the Rutronik System Solutions demonstration is the PSOCTM Edge E84 and its corresponding evaluation kit—acting as the “brain” of the system. AI applications typically rely on cloud computing to handle vast amounts of data. However, by integrating the PSOC Edge E84, this processing power is brought to the edge, creating a self-contained system. This transition not only enhances performance but also strengthens security through advanced features such as the lockstep-secured enclave in a low-power, always-on domain, Infineon’s Edge Protect Category 4, and standard Trusted Firmware-M activation with Mbed-TLS for cryptographic operations.
High-performance microcontrollers for AI applications
The PSOCTM Edge E84 series, built on Arm Cortex-M microcontrollers, delivers high-performance, secure, and energy-efficient processing with hardware-supported machine learning acceleration. These devices are powered by the Cortex-M55, featuring Helium DSP support and Arm Ethos-U55 integration, alongside a Cortex-M33 incorporating Infineon’s ultra-low-power NNLite hardware accelerator. This enables efficient neural network processing, making them ideal for AI and machine learning applications.
Enabling vision through radar technology
The demonstration system “sees” using the Rutronik Adapter Board RAB3 – Radar, which connects to the evaluation kit via an Arduino interface. This board features Infineon’s XENSIV 60GHz radar sensor, integrating cutting-edge radar technology into the system. Radar sensors offer unique advantages, such as the ability to function behind certain materials, rendering them invisible while maintaining hygienic, easy-to-clean surfaces. This makes them particularly suitable for environments with strict hygiene requirements. Additionally, radar technology excels in outdoor applications, operating reliably even in direct sunlight. The XENSIV sensor’s compact form factor and low power consumption, combined with its L-shaped antenna array, enable comprehensive horizontal and vertical angle measurements, ensuring accurate radar scanning.
Accelerating development with pre-trained AI models
Training AI models for gesture and voice recognition is a time-intensive process, but it is critical for system functionality. Rutronik System Solutions has pre-trained the microcontroller to recognise a range of hand gestures and predefined voice commands (Down, Go, Left, No, Right, Stop, Up, and Yes). To improve responsiveness to imprecise gestures, the system currently operates without a neural network at this stage. However, if greater robustness is required, a neural network can be integrated to enhance classification accuracy.
Moreover, the PSOC Edge E84 includes always-on acoustic activity detection, facilitating HMI operation with minimal power consumption. This is particularly beneficial for battery-operated devices, extending their operational lifespan. The demonstration system integrates Infineon digital microphones, acting as its “ears” to enable near real-time voice control of the motor.
Seamless motor and display integration
Motor control is implemented via an eight-pole BLDC motor with integrated Hall sensors, also from Infineon. This motor runs in parallel with speech recognition and gesture control processes on the same microcontroller. Complementing the system is a 1024 x 600 IPS TFT LCD touch display from Raystar. Supported by the PSOC Edge E84’s 2.5D graphics processing unit and MIPI-DSI/DBI interface (up to 1024 x 768), this display ensures high-quality visuals with optimised power consumption.
Optimising AI for Edge computing
By shifting AI processing from the cloud to the edge, Rutronik System Solutions is enabling significant advancements in smart applications, particularly for outdoor environments. “Our demo systems, enhanced by radar technology, are far less susceptible to sunlight interference, offer greater data security with an ‘on the edge’ approach, and significantly reduce latency—a major drawback of traditional cloud-based solutions,” explains Stephan Menze, Head of Global Innovation Management at Rutronik. “With this demonstration—like all of our developments—we integrate the most advanced components from our partners, allowing customers to create smart applications in a fraction of the typical development time.”
