Validate core functions and operating conditions step by step before field deployment.
Robowink’s PoC is not a simple demonstration. It is a technology-validation process that checks AMR operating conditions and the applicability of intelligent safety technologies in real environments.
The PoC reflects real site conditions, including travel routes, movement of workers, carts and forklifts, sensor installation position, network conditions, and control interfaces. Collected logs and test results are then used to define the deployment scope and improvement items.
AMR operation and the DVS+SNN-based Safety Assist function are validated separately, followed by system-level integration validation.
We verify route feasibility, bottleneck areas, turning and stopping space, and repeated-operation conditions on the actual route.
We verify event-data quality and sensor-installation conditions in areas with frequent motion.
We evaluate approach-situation classification and decision stability using event patterns that change over time.
We verify that perception results integrate as assistive information without conflicting with the AMR’s existing navigation and safety systems.
Current field validation uses commercially available development platforms to acquire data quickly and validate algorithms and interfaces.
Instead of a one-time demonstration, we define repeatable test conditions and compare results under the same conditions.
Establish baseline logs and normal-state behavior during unobstructed operation.
Observe event changes when a worker or cart enters from the side.
Compare perception and decision logs according to approach direction and relative motion changes.
Evaluate perception stability in complex situations where multiple moving objects cross paths.
Analyze error characteristics when objects are partially occluded or event density becomes concentrated.
Specific acceptance criteria are defined according to the site and PoC scope, with measurements centered on the following items.
Classification accuracy, false positives / missed detections, and result distribution by scenario
Latency from sensor input to decision and interface output
Result consistency and system status during repeated tests and long-duration operation
Suitability of sensor position, field of view, communications, routes, and integration with the existing AMR
※ These figures are current R&D and PoC targets and do not represent guaranteed field results or product specifications. Test criteria may vary according to the application environment and system configuration.
We define site requirements first, agree on test scenarios and evaluation methods, and then proceed with the PoC.
Define the target process, routes, problem situations, and target KPIs
Check aisles, sensor installation, network, and safety conditions
Define the PoC configuration and repeatable test scenarios
Collect data, event, and state logs under actual operating conditions
Analyze failure cases and adjust thresholds, models, and installation conditions
Document validation results, limitations, and deployment / scale-up recommendations
To support a deployment decision, we provide deliverables that include test conditions and data rather than only demonstration results.
Share your application environment and the problem situations you want to validate, and we can help define the test configuration and validation items.
