Specifications
NPU
Mobilint REGULUS AI SoC — 10 TOPS (INT8)
CPU
Quad-core Arm Cortex-A53 up to 1.2 GHz + Cortex-M0+
Memory Type
LPDDR4 / LPDDR4X
Memory Capacity
8 GB (4 GB optional)
Storage
32 GB eMMC (16 GB optional) + microSD x1
Operating System
Linux 5.15
Thermal Design Power
10 W (TBD)
Ethernet
2 ports
Serial
RS-232/422/485 x1
CAN
CAN FD x1
Digital I/O
DI x4 / DO x4
USB
USB 3.1 Gen1 x4
Camera Inputs
MIPI CSI x2 (4-lane) + USB UVC + IP cameras
Integrated ISP and Video Codec
Yes — camera preprocessing and inference both run inside REGULUS
Integrated Codec Path
Camera input, preprocessing and inference on one SoC — no external accelerator
Model Coverage
490+ open and closed-source models validated on the NPU architecture
Mounting
Vehicle / machine mount for AGVs, AMRs and forklifts
Price
Quote upon request
Availability
Available
Overview
The MLX-R1 mounts directly onto forklifts, AGVs and AMRs and runs AI entirely on-device using Mobilint's REGULUS AI SoC. REGULUS integrates a quad-core Arm Cortex-A53 CPU cluster, a Cortex-M0+ real-time core, the NPU, an ISP and a video codec on a single die, so camera input, preprocessing and inference all happen inside one chip with nothing external required.
I/O is oriented to mobile machinery: two Ethernet ports, RS-232/422/485, CAN FD, four digital inputs and four digital outputs, four USB 3.1 Gen1 ports, and two 4-lane MIPI CSI camera inputs alongside USB UVC and IP camera support. The system runs Linux 5.15 with 8 GB of LPDDR4X (4 GB optional) and 32 GB of eMMC (16 GB optional) plus microSD.
The REGULUS NPU architecture has been validated across more than 490 open- and closed-source models, covering vision, voice and language workloads, so forklift safety monitoring, pedestrian and hard-hat detection, restricted-zone intrusion detection and collision warnings all run without a network connection.
QS Compute supplies the MLX-R1 together with the MLM-1 SoM, REGULUS AI EVK and MLD-R1 USB accelerator for teams that need to develop on the SoC and then deploy inside an enclosed, vehicle-mounted system.
Key Benefits
Complete AI pipeline in one box — ISP, codec, CPU and NPU on a single SoC. CAN FD and dual Ethernet for direct machine integration. Fully on-device inference for sites with no connectivity. Validated across 490+ models. ~10 W envelope for battery-powered mobile equipment.
Applications
AGV, AMR and forklift safety and navigation; construction and mining machinery hazard detection; industrial worker-safety monitoring; logistics-centre vision; mobile terminals in warehouses; agricultural machinery perception.
Request a Quote — MOBILINT MLX-R1 MINI AI BOX — REGULUS EDGE AI COMPUTER FOR VEHICLES
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →