This blog post was originally published at Infineon’s website. It is reprinted here with the permission of Infineon.
Across home, retail, and industry, devices that once followed simple rules are now expected to understand people and context. A thermostat shouldn’t just follow a schedule; it should know if anyone is in the room and choose the preferred setting for the identified people in the room. A POS terminal shouldn’t just show a keypad; it should adapt to environmental conditions, evolving security threats, and protocols. An appliance shouldn’t just flash an error code; it should guide the user through what to do next.
This marks a shift from connected to aware and adaptive. PSOC™ Edge MCUs, combined with Infineon’s 60 GHz XENSIV™ radar, AIROC™ CYW55513 Wi-Fi and Bluetooth® combo, HYPERRAM™, NOR flash, PDM/analog microphones, humidity and temperature sensing, camera support, and DEEPCRAFT™ AI software, are designed for exactly these devices.
From schedules to presence‑driven control
Many systems still rely on timers and simple motion sensors. They heat and cool based on a clock, not real usage. They run full power cycles even when nobody is around.
With presence and environmental sensing on PSOC™ Edge, that model changes. A room controller can observe who is there, how conditions evolve, and when spaces are consistently empty. Instead of a static schedule, you get a thermostat or HVAC HMI that:
- Wakes its display only when someone approaches
- Pre‑heats or cools just before typical occupancy
- Identifies people in the room and understands what type of setting these people use
- Automatically shifts to energy‑saving modes when rooms are unused
Appliances benefit from the same awareness. Air purifiers, range hoods, and AC units can adjust airflow and power dynamically based on occupancy and humidity, rather than running fixed programs. The device responds to the way spaces are used, not just to a setpoint.
From touch‑only to multimodal interaction
Once devices understand presence, interaction can evolve. Many environments make pure touch suboptimal: hands are full in a kitchen, gloved on a factory floor, or contact‑sensitive in healthcare. PSOC™ Edge lets you layer voice and gesture on top of traditional touch:
- A POS terminal can still present a familiar touchscreen, but add radar‑based “next/back/confirm” gestures to complete simple flows without contact
- An oven or dishwasher can keep its front panel, yet respond to “start,” “pause,” or “add 5 minutes” spoken locally—even when the user can’t reach the controls
- A smart lock can offer a compact keypad or UI, wake automatically as someone approaches, and accept a simple gesture to confirm a choice, or identify the individual and allow access
The interaction model becomes flexible: touch when convenient, voice when hands are busy, gesture when hygiene or distance matters, and identification when required.
From one‑size‑fits‑all UX to context‑aware guidance
When sensing meets multimodal interaction, HMIs can do more than show static menus.
With local ML on PSOC™ Edge, content adapts to context:
- A thermostat or HVAC HMI can move from cryptic error codes to clear, step‑by‑step guidance when sensors detect a probable issue—such as a clogged filter or abnormal runtime
- Wearables and handhelds can use IMU data and camera input to adjust overlays—simplifying the interface during fast movement and surfacing detailed instructions when the user is stationary and focused on a specific object
- A robot’s compact HMI can show live operating data most of the time, but switch to a guided diagnostic overlay when sensor patterns indicate a fault
Instead of every user needing to know what to do, the device explains itself at the right moment.
From cloud‑dependent to reliable, local intelligence
The more devices sense and adapt, the more decisions they make. Doing all of this in the cloud leads to latency, bandwidth costs, and fragile behavior when connectivity is weak. Processing data locally not only cuts latency and bandwidth but also protects user privacy—raw sensor data stays on the device unless explicitly authorized, reducing exposure and easing regulatory compliance.
With PSOC™ Edge and DEEPCRAFT’s optimized models, key intelligence runs locally so the device can perform core functions without a cloud connection:
- Voice: frequently used commands and wake words are processed on‑device, keeping basic functionality responsive and private – even when offline; the cloud is used only for complex intents or updates
- Sensing and vision: presence detection, simple classification, and anomaly detection operate at the edge, sending only meaningful events upstream
The cloud remains a partner for updates, analytics, and coordination—while day‑to‑day interaction and safety‑critical decisions stay anchored in the device.
Try it now: the PSOC Edge E84 HMI Kit
These behaviors aren’t just conceptual—they’re designed to be hands‑on with the PSOC™ Edge E84 HMI Kit. On a single kit, you get:
- A 4‑inch 480×480 capacitive touchscreen that maps naturally to a thermostat, appliance front panel, compact POS, or access UI
- A microphone array, 60 GHz radar, and a USB camera to combine voice, touchless gesture, presence detection, and basic vision on the same PSOC™ Edge MCU
- Built‑in temperature, humidity, and IMU sensors, plus micro-SD card and mikroBUS™ expansion interface to explore environment‑aware and motion‑aware scenarios
What you can prototype in a weekend
- A presence‑driven intelligent thermostat that wakes on approach, adapts to real occupancy, and responds to voice
- A smart POS or kiosk that supports touchless navigation and context‑aware prompts
- A smart appliance front panel that offers guided flows and hands‑free control
- A smart lock or access panel that anticipates interaction and keeps core logic local
Developer experience, security, and privacy
- Build fast with ModusToolbox™, UI frameworks like LVGL, and example projects that integrate radar, audio, and camera
- Deploy models with DEEPCRAFT™’s optimized pipelines for wake words, gesture classification, and anomaly detection
- Keep trust front‑and‑center with hardware security features such as secure boot, crypto accelerators, and protected credentials—plus local voice processing to avoid sending raw audio off the device to eliminate privacy risks
- Tell us what you’re creating: presence‑aware thermostats, touchless POS, guided appliances, access panels, or something creative with Edge intelligence
- Share tips, ask questions, and request sample code or demos—your feedback helps shape future releases
- If you’ve tried the kit, post your measurements, UX ideas, or integration notes with radar, audio, and camera
If you want a quick‑start checklist or links to demos and reference projects, comment below or ask questions and share your ideas and projects on the PSOC™ Edge community page!
