Innovation
Robotics in Innovation: articles, explanations and technology developments with context and verified sources.
ISO 9283 provides criteria and test methods for evaluating different characteristics of industrial manipulating robots. Knowing which parameter was measured—and what information is missing—helps compare specifications without treating a figure as a guarantee for every task or installation.
The label “collaborative” does not, by itself, prove that an installation is safe. An assessment must consider the robot, the task, the tools and the working environment.
Battery runtime is not enough to compare mobile robots. Energy per distance or per task may be more informative, but depends on how it is measured, the payload and the environment. The documentation verified in this review does not confirm a common method for comparing AMR energy consumption.
A 3D scene or simulation program is not enough to prove that a digital twin connected to a robot exists. The key is to document which physical entity it represents, what information it exchanges with that entity, how the model is updated, and what evidence validates the correspondence.
ROS 2 offers tools for organizing execution, but using it alone does not establish that a robot meets timing deadlines. Evaluation also depends on the operating system, hardware, application, and representative testing.
Regulation (EU) 2023/1230 does not classify all robots in the same way. To guide a review, identify the product and consult the applicable legal text: calling something a “robot” is not enough to establish which requirements apply.
A score describes the outcome of a particular test, not a robot’s universal capability. To judge whether it applies beyond the lab, examine the tasks, environments, hardware, metrics, and evaluation conditions.
The safety of a mobile robot cannot be demonstrated with a sales datasheet. This guide explains what documentation to request, how to define intended use, and what to validate in the actual installation before authorising operation.
A robot designed for collaborative applications does not prove that a particular installation is safe. Before putting it into operation, assess the complete system, document its protective measures, and check them under real working conditions.
A single demonstration is not enough to establish autonomy. Assessing a claim requires a defined task, explicit test conditions, and metrics that make successes, errors, and human intervention understandable.
The IFR counted more than 600,000 industrial robot installations in 2025 and an operational stock of five million. The figures describe real expansion, but on their own they do not measure productivity, employment or the tasks being automated.
An AMR is not validated by a navigation demonstration. Assessing a deployment requires applicable requirements, safety tests in the actual environment, and documented operational integration.
ManipBench evaluates vision-language models on reasoning tasks related to manipulation. Its results help measure specific capabilities, but do not certify that a robot can work autonomously and reliably in every environment.
The label “collaborative” does not, by itself, show that a particular task is safe. Evaluation must address the actual application, its tasks and its configuration; NIST research provides study methods, but does not certify a particular cell.
The draft attributes a scope, publication date and safety exclusion to ISO 18646-5:2026. The verified sources available do not document this standard, so they cannot confirm those claims or explain its tests.
