Automated robotics platforms have fundamentally changed warehousing logistics, reducing manual labor during 24/7 operations while enhancing operational efficiency. However, the industry continues to advance: While traditional automated guided vehicles (AGVs) rely on fixed paths, autonomous mobile robots (AMRs) offer safer, more dynamic navigation for greater deployment flexibility in constantly changing environments. This is achieved by combining modern artificial intelligence (AI) capabilities with advanced sensor arrays. Since AGV-reliant companies have already invested heavily in powerful robotics platforms, retrofits provide a cost-effective way of upgrading existing infrastructure. This blog discusses AGV-to-AMR upgrades before introducing the SECO SOM-SMARC-QCS6490 computer-on-module (COM) as a forward-thinking solution for next-generation warehouse automation.
Automated robotics platforms have revolutionized warehousing and logistics applications. In particular, automated guided vehicles (AGVs) have enabled companies to enhance operational efficiency through mechanized stock transportation, reducing manual labor requirements for 24/7 operations. However, AGVs long predate modern robotics: In the 1950s, AGVs consisted of modified towing tractors guided by the magnetic field of floor-embedded wires. Since then, AGVs have evolved into sophisticated robotics platforms that typically use optical, laser, or magnetic-tape-based guidance systems. Still, these devices are limited to movement along predetermined paths based on fixed guidance references.
Autonomous mobile robot (AMR) technology offers a way to overcome this evolutionary quirk, leveraging edge artificial intelligence (AI) and advanced sensor arrays to support open-environment navigation. For logistics, this significantly boosts operational flexibility and enables dynamic obstacle negotiation to enhance safety for humans working alongside autonomous equipment. But despite the benefits, many companies may be resistant to AMR adoption because they have already invested millions in AGVs, having equipped facilities with AGV infrastructure.
This is where retrofits come in. Modern AGVs already integrate advanced robotics systems for precise movement in warehouse settings. By keeping this foundation and adding new sensing and processing systems, companies can convert their existing AGVs into AMRs, accelerating time-to-market while retaining equipment investments.
Changing an AGV’s mind: New brain, new game
AGVs are advanced machines, but their on-board intelligence and perception capabilities are basic compared to AMRs. To release existing AGV platforms from fixed navigation paths, the first step is to retrofit them with a new “brain” that supports advanced edge intelligence. The “brain” must then be connected to a wider “nervous system” consisting of additional sensors such as 3D lidar modules, cameras, and inertial measurement units (IMUs). In combination, these systems enable an AGV to perceive the natural features of its surroundings, instead of set path markers. Sensor fusion processing then enables advanced open-environment navigation, as well as people or obstacle detection for safer operation.
In addition, an AGV-turned-AMR can log surroundings to support simultaneous localization and mapping (SLAM). In many cases, existing wheel odometry data streaming from the original AGV system can be combined with new sensor data for even better positioning and distancing accuracy within large warehouse facilities.
AMRs also benefit from a connection to overarching fleet orchestration platforms. Usually hosted in the cloud, these provide warehouse managers with granular visibility into fleet performance so that operations can be adjusted to boost efficiency. For AMR technical teams, telemetry reports can then support predictive maintenance efforts to maximize fleet uptime.
While AGV-to-AMR retrofits can reduce expenditure by repurposing existing investments, it is worth acknowledging the burden of hardware compatibility and software development for retrofit designers. However, commercial-off-the-shelf (COTS) solutions can help by providing ready-to-go embedded computing platforms that allow developers to focus on AGV retrofit success, rather than combining a solution’s individual components.
Accelerating AGV-to-AMR retrofit development using COTS ecosystems
When designing retrofits, developers often require compact solutions that offer low power consumption. Developers can then fit upgraded solutions within target platform enclosures while using the existing power infrastructure.
For AGV-to-AMR conversions, the SMARC (smart mobility architecture) open computer-on-module (COM) standard presents an ideal solution for brain retrofits. It offers a wide range of interfaces in an 82- by 50-mm form factor with a narrow power envelope, typically under 6 W. Furthermore, open COMs, also known as system-on-modules (SOMs), benefit from long-term availability and interoperability between solutions, reducing the chance of vendor lock-in. COMs are typically mounted on application-specific carrier boards, enabling AGV-to-AMR retrofit designers to upgrade intelligence via modular substitution as chipsets evolve.
An example of such a solution is SECO’s SOM-SMARC-QCS6490, which is based on the Qualcomm Dragonwing QCS6490 processor. It provides AGVs with the following intelligence boost to support AMR functionality:
- Multi-core CPU: The QCS6490 offers 8x application cores (1x Kryo Prime up to 2.7 GHz, 3x Kryo Gold at 2.4 GHz, 4x low-power Kryo Silver at 1.9 GHz—all based on Arm architecture) for platform control and simultaneous data processing from many sensors.
- Adreno GPU: The dedicated GPU enables advanced on-device HMIs and efficient multi-camera data ingest for autonomous off-path navigation.
- Embedded AI accelerator: Efficient on-processor AI at up to 12.15 INT8 TOPS supports low-latency inferencing for safe open-environment navigation and real-time visual- and lidar-based SLAM—in combination with the main CPU.
- Flexible OS support: Retrofits benefit from compatible integration and synchronization with established operational frameworks. The SOM-SMARC-QCS6490 supports Windows 11 IoT Enterprise, Yocto Linux, and Android for easy adoption.
Alongside the QCS6490 processor, the SOM-SMARC-QCS6490 features up to 12 GB soldered-down LPDDR5-6400 memory for reliable high-volume data processing within challenging operating environments. This is further enhanced by an extended temperature version (-30°C to +85°C) that enables AGV-to-AMR retrofits across a full range of warehousing climates, including cold storage.
Like other SMARC modules, the SOM-SMARC-QCS6490 features a wide set of interfaces, including USB and MIPI CSI for sensor and camera arrays, optional Wi-Fi for cloud connectivity to orchestration platforms, and CAN bus for AGV motion system integration. For developers looking to test these interfaces, the SECO DEV-KIT-SMARC provides a ready-to-use, industrial-grade carrier system for accelerating development with SECO’s SMARC modules. Together with physical ports and interface headers for subsystem connections, a wide voltage input range of 9 to 20 V supports direct retrofit prototyping with existing AGVs.
Furthermore, the SOM-SMARC-QCS6490 is supported by SECO’s Clea software framework, which includes the modular, Yocto-based Clea OS, as well as Clea Portal and Clea Edgehog for AMR fleet monitoring and device management. Accelerating software development for AMR retrofits, the Clea software ecosystem supports essential functions like secure boot and OTA updates to promote long-term compliance and success for AMR fleets. Its DevOps framework, along with supporting SBOM and other collateral, facilitates cybersecurity certifications, such as those required by the European Union’s RED Directive.
Conclusion
By repurposing existing robotics infrastructure, AGV-to-AMR retrofits potentially save significant investment and development costs to warehousing businesses and warehousing solution OEMs. In combination with advanced sensor arrays and SECO’s Clea software framework, the SOM-SMARC-QCS6490 enables heightened environmental perception for autonomous open-environment navigation in unpredictable warehouse settings. Furthermore, SECO provides an evaluation platform that supports comprehensive retrofit testing with real AGVs, accelerating time-to-market for an AMR technological refresh in line with the open COM philosophy.
To explore SECO’s full range of hardware and software solutions for industrial robotics, head to seco.com.