Electronica 2024 created the perfect opportunity to showcase Alif’s cutting-edge solutions for AI, endpoint devices, and embedded systems. Our booth was bustling with engineers, members of the technical press, and industry professionals, who saw demos of how Alif’s products solve real-world problems. Henrik Flodell, Alif’s Senior Director of Product Marketing, delivered a captivating presentation about the definition of AI. Let’s recap the highlights of this year’s show. 

Alif films a press video at electronica 2024.

NVIDIA TAO Demo 

The NVIDIA TAO Demo was undoubtedly a crowd favorite. Leveraging NVIDIA’s TAO Toolkit, the demo featured a license plate-detection system powered by the Alif Ensemble MCU. This system combined a MobileNetV2 backbone with a YOLOv3 detection head to demonstrate high inference speeds and low power consumption. The model, trained on a robust dataset using Edge Impulse, achieved inference times of 10 ms with INT8 precision on the Ethos-U55 neural processing unit (NPU).

The solution effectively showcased our ability to deliver high-performance edge AI that minimizes power consumption. Combining streamlined edge inference with robust scalability, the demo was lauded by engineers for its ability to reduce latency, operational costs, and power requirements.

Visitors streamed into Alif’s electronica booth.

Balletto Auracast Demo

Another star was the Balletto Auracast Demo, which showcased the unmatched performance of Alif’s Balletto MCU in BLE audio applications. Specifically, this demo highlighted real-world applications, such as health monitoring, fitness trackers, and advanced wearables. With LC3 encoding powered by the Cortex-M55 and Helium vector engine alongside the Ethos-U55 NPU, the Balletto platform demo achieved encoding speeds 2.5 times faster than conventional MCUs, all while maintaining ultra-efficient energy consumption.

Attendees saw firsthand how Balletto pushes technical boundaries and accelerates time-to-market for developers in the wearables and IoT industries. 

Other Notable Demos

Beyond the NVIDIA TAO and Balletto Auracast demos, Alif Semiconductor showcased a suite of cutting-edge demonstrations that emphasized the versatility and power of its platforms:

• Nota-AI In-cabin safety: Using Alif’s Ensemble E7 AppKit, this demo can detect if a driver is drowsy, distracted, falling asleep at the wheel and more, to keep drivers safe.
• OpenMV Cam AE3: Measuring at 1 inch by 1 inch and using the dual core Alif E3 processor, this tiny camera processes over 100 frames per second at  <100mA total system power draw.
• Face detection and identification: Running on Alif’s AI/ML AppKit, a full featured platform for edge-based AI/ML development, this demo highlighted face detection and identification using only 0.8 uJ of energy per inference operation. The system can be trained to remember the landmarks of different faces, and will automatically highlight a face that it has previously been instructed to remember.  
• License Plate-Detection System: In addition to the NVIDIA TAO-powered solution, a real-time decoding system demonstrated character recognition with low latency on Alif’s Ensemble MCU.

Insights on AI and the Future

Another highlight came when Flodell delivered an insightful presentation at the Embedded Forum, organized by Embedded.com. His talk, “What is the Definition of AI and What Does This Mean for the Future?” examined the divergence between AI definitions—human-like intelligence versus statistical problem-solving—and their convergence in embedded applications. 

Henrik Flodell presents at the Embedded Forum with Embedded.com at electronica 2024.

Flodell explored the major differences between cloud-hosted AI and embedded endpoint AI: the challenges of energy efficiency, data privacy, and computational constraints. These distinctions form the foundation of Alif’s design philosophy, which prioritizes localized intelligence to reduce latency and improve operational security.

Flodell also discussed real-world examples, from healthcare devices to autonomous robotics, to demonstrate the transformative impact of edge AI. By combining energy-efficient hardware with secure processing capabilities, Alif is enabling innovative applications previously considered impractical in embedded environments.

The AI Takeaways

Beyond a technology showcase, electronica 2024 was a success because it created an opportunity to demonstrate our products’ relevance and impact. Engineers left with a clear sense of how Alif’s solutions enable smarter, faster, and more efficient embedded systems while addressing critical concerns like cost, power, and scalability. As industries increasingly demand AI-powered embedded solutions, Alif’s technology is paving the way for a future where devices are more capable, connected, and efficient.