Products

SoM, Accelerator & Connectivity

System On Module

Physical AI Multi-modal

Edge Physical AI

Multi-modal Physical AI platform with vision, audio, and environmental sensors. On-device inference with real-time actuation for autonomous systems.

AI Runtime: TensorFlow Lite / ONNX / TinyML

Sensors: Camera (MIPI CSI), microphone, IMU, env

Compute: NPU + MCU co-processing

Power: Ultra-low-power for battery operation

Application: Robotics, smart wearables, safety systems

Remote Unmanned Vehicles

Edge Remote

Plug-in Edge AI SoM for remote unmanned vehicles — adds autonomous perception and on-device intelligence without modifying the existing control system. Adaptable to aerial, surface, and underwater platforms.

AI: NPU-accelerated real-time on-board inference

Connectivity: 5G / WiFi 6 multi-link, acoustic option for underwater use

Integration: Non-intrusive add-on to existing vehicle platforms

Enclosure: Ruggedized options for harsh and high-pressure environments

Application: Autonomous drones (UAV), underwater vehicles (AUV), unmanned surface vessels

Edge AI / Physical AI

Adaptive Edge AI SoM

BSP-ready system-on-module designed for Edge AI and Physical AI applications, combining embedded processing with configurable AI acceleration for rapid product development.

Compute: Arm-based processing / FPGA-configurable acceleration

Connectivity: 5G / Wi-Fi 6 / BLE / Ethernet

OS: Linux / RTOS / Android

BSP: Pre-integrated, partner-ready

Applications: Drones, robotics, industrial vision, smart sensing, AI wearables

Accelerator

Reconfigurable Compute Accelerator

Reconfigurable Computing

FPGA-based acceleration platform for prototyping, networking, hardware acceleration, and ASIC pre-validation.

FPGA: AMD FPGA ecosystem compatible

Interface: M.2 PCIe / PCIe x4

Use Cases: Algorithm prototyping, networking acceleration, hardware acceleration, ASIC pre-validation

Development: RTL / HLS support

Connectivity

Sensor Hub modules designed to aggregate multi-sensor data at the edge — enabling always-on, low-power sensing with wireless connectivity for Physical AI systems.

Always-On BLE Sensing

Ultra-Low Power Sensor Hub

Ultra-low-power sensor hub for always-on, battery-powered sensing. Aggregates environmental and motion sensor data over BLE — purpose-built for long-life IoT deployments.

Core: Ultra-low-power MCU optimized for always-on operation

Connectivity: BLE 5.3 wireless

Sensors: Environmental, motion, IMU interface

Power: Sub-mW standby — coin-cell compatible

Application: Smart buildings, asset tracking, wearable sensing

AI-Enabled Sensor Hub

AI Sensor Hub

Sensor hub with on-device NPU inference. Processes and filters multi-sensor data locally before wireless transmission — reducing bandwidth and enabling edge AI decisions at the sensor layer.

Core: Low-power MCU with integrated NPU for on-device AI

Connectivity: BLE 5.3 + optional multi-protocol

AI: On-device sensor data inference and anomaly detection

Sensors: Multi-sensor fusion (environmental, vision, motion)

Application: Industrial monitoring, predictive maintenance, smart wearables

Wireless Bridge & Relay

WiFi Relay Module

Compact WiFi relay module for extending wireless coverage and bridging sensor networks to edge AI systems. Enables mesh connectivity in environments with limited direct line-of-sight.

Protocol: WiFi 6 / WiFi 5 relay and bridging

Interface: BLE-to-WiFi / sensor-to-cloud bridging

Topology: Mesh network support for extended coverage

Power: Low-power relay operation

Application: Factory floor sensor networks, drone ground relay, aquatic surface relay

Application

AI application solutions built on LEAPSWing SoM platforms — ready-to-integrate on-device inference engines.

On-Device Speech AI

Voice Recognition

On-device voice recognition and speech AI solution. Keyword detection, command recognition, and SLM-based natural language processing — fully offline, privacy-preserving.

Models: Keyword spotting, ASR, SLM (Small Language Model)

Runtime: NPU-optimized speech pipeline — low latency

Input: Microphone array / I2S audio interface

Privacy: On-device processing — no audio transmitted to cloud

Application: Smart toys, service robots, industrial HMI, voice-controlled wearables