Robot Obstacle Avoidance Function Testing: Standards, Methods, and Technical Requirements

Robot obstacle avoidance function testing is a critical safety and performance verification process for autonomous mobile robots, AGVs, service robots, and household cleaning robots. This testing evaluates whether a robot can reliably detect static and dynamic obstacles, execute protective stops, and perform avoidance maneuvers in accordance with international standards such as ISO 3691-4, ISO 13482, and IEC 60335-2-2. In this article, we provide a comprehensive technical overview of obstacle avoidance testing samples, test items, methodologies, and typical applications to help manufacturers prepare for compliance evaluation at accredited third-party laboratories.

Technical Overview

Obstacle avoidance is one of the most fundamental safety functions of any mobile robotic system. Whether the robot operates in a warehouse, a hospital corridor, a private lawn, or a living room, its ability to perceive its surroundings and react before physical contact occurs directly determines the safety of nearby persons, the robot itself, and the surrounding property. Robot obstacle avoidance function testing is the systematic process of verifying that the perception system, control logic, and actuation response of a robot perform as designed under repeatable, standardized conditions.

Modern obstacle avoidance systems typically rely on a fusion of multiple sensing technologies, including LiDAR (light detection and ranging), ultrasonic sensors, infrared proximity sensors, time-of-flight (ToF) cameras, structured light depth cameras, stereo vision systems, and bumper contact switches. Each sensing modality has distinct strengths and weaknesses. For example, LiDAR offers precise ranging but may struggle with highly absorbent, transparent, or specular reflective surfaces, while ultrasonic sensors handle glass reasonably well but provide lower angular resolution. Because of these limitations, functional testing must cover not only nominal operating conditions but also edge cases such as dark environments, low-contrast obstacles, glass panels, mirrors, and small-diameter objects.

Several international and national standards govern obstacle avoidance evaluation depending on the robot category:

• ISO 3691-4 – Safety requirements for driverless industrial trucks and their systems, defining test bodies, protective separation distances, and speed control requirements for AGVs and AMRs.
• ISO 13482 – Safety requirements for personal care robots, including mobile servant robots and physical assistant robots.
• ISO 10218-1/-2 – Safety requirements for industrial robots and robot systems, relevant to mobile manipulators.
• ISO 18646 series – Performance metrics for service robots, covering obstacle detection and avoidance capability benchmarks.
• IEC 60335-2-2 – Safety requirements for household and similar electrical appliances, including robotic vacuum cleaners, with clauses addressing obstacle encounter and hazard avoidance behavior.
• GB and GB/T national standards – Chinese national standards for service robots, electrical safety, and electromagnetic compatibility that apply to products sold in the Chinese market.
• UL standards – North American safety requirements applicable to consumer and commercial robotic products.

Testing is typically performed by accredited third-party laboratories to support safety certification, regulatory compliance, product development validation, and market access documentation.

Testing Samples

Autonomous Mobile Robots (AMRs)

AMRs used in warehousing, manufacturing, and logistics are tested for navigation-grade obstacle detection at operational speeds. These robots frequently employ 2D safety LiDAR and 3D depth sensing, and they must demonstrate reliable protective stopping and route re-planning when persons or objects enter their path. Test emphasis is placed on approach speed, braking distance, and detection of low-profile obstacles.

Automated Guided Vehicles (AGVs)

AGVs operating on fixed or semi-fixed routes are evaluated against ISO 3691-4 requirements, including detection of standard test bodies of various diameters, speed reduction zones, and stopping performance on ramps and in narrow aisles.

Household Cleaning Robots

Robot vacuum cleaners and floor mopping robots are tested per household appliance safety requirements, focusing on obstacle encounters, stair-edge (cliff) detection, entanglement avoidance, and safe behavior around cords, curtains, and pet water bowls.

Service and Delivery Robots

Delivery robots, hotel service robots, disinfection robots, and companion robots are assessed for human detection, protective stop performance, and graceful avoidance in shared pedestrian spaces under both static and dynamic obstacle conditions.

Outdoor and Special-Purpose Robots

Robotic lawn mowers, inspection robots, and agricultural robots require additional evaluation under sunlight interference, uneven terrain, dust, and rain, with special attention to boundary and object detection performance.

Testing Items

Typical test items in an obstacle avoidance evaluation program include the following:

• Static obstacle detection: verification that the robot detects and stops or steers around stationary objects of defined shapes, sizes, and surface properties.
• Dynamic obstacle detection: evaluation of detection and response when obstacles move toward or across the robot path, including approaching pedestrians and crossing vehicles.
• Human detection and protective stop: confirmation that the robot reliably detects test dummies or standard human-representative test bodies and comes to a controlled stop before contact.
• Stopping distance measurement: quantification of the total stopping distance from obstacle detection to complete standstill at various travel speeds and payload conditions.
• Avoidance success rate: statistical measurement of successful avoidance events over repeated trials, typically expressed as a minimum success percentage over a defined number of runs.
• Response time: measurement of sensor-to-actuation latency, including detection delay, control processing time, and brake engagement time.
• Low-lying and thin obstacle detection: testing with objects such as cables, socks, small cylinders, and poles of small diameter (for example, 70 mm test bodies referenced in ISO 3691-4).
• Transparent and reflective object detection: evaluation against glass panels, mirrors, and polished metal surfaces that challenge optical sensors.
• Dark and low-contrast environment testing: verification of performance in low illuminance and against obstacles with colors similar to the floor.
• Cliff and stair-edge detection: for indoor robots, confirmation that downward-facing sensors prevent falls from elevated surfaces.
• Bumper and contact response: validation of mechanical bumper activation, contact force limits, and post-contact stopping behavior when avoidance fails.
• Recovery and re-navigation behavior: assessment of how the robot resumes its task after an avoidance event, including escape from entrapment or dead-end configurations.

Testing Methods

Testing is conducted on a standardized test field with controlled floor properties, controlled lighting, and instrumented measurement equipment. High-speed cameras, laser distance meters, timing gates, and force sensors are used to capture objective data on detection points, stopping positions, and contact events.

Standard Test Bodies and Obstacle Configurations

Standards such as ISO 3691-4 define cylindrical test bodies with specific diameters (for example, 70 mm, 200 mm, and 400 mm) and matte black or specified surface finishes to represent legs, persons, and bulkier obstacles. These test bodies are positioned vertically or horizontally on the robot path, and the robot approaches them at defined speeds. The test verifies that the protective devices detect the obstacle in time to stop or maintain a safe separation distance before contact.

Approach and Crossing Scenarios

The robot is driven toward obstacles at maximum operational speed, reduced speed, and creeping speed. Dynamic tests introduce obstacles moving toward the robot or crossing its path at representative pedestrian speeds. Contact velocity, if any contact occurs, is measured and compared against allowable limits defined by the applicable standard.

Speed and Braking Distance Measurement

Stopping distance is measured from the point at which the protective device initiates the stop command to the point of complete standstill. The measured values are compared with the calculated protective separation distance, which depends on robot speed, response time, braking capability, and floor conditions. Tests are repeated on level surfaces and on the maximum rated slope of the operating environment.

Environmental Variation Testing

Because sensor performance varies with environmental conditions, test sequences include low-light operation, strong ambient light or sunlight exposure, dust or particle presence, and surfaces with different reflectivity and texture. Transparent, mirrored, and black absorbent obstacles are included to stress the limits of optical sensing.

Repetition and Statistical Evaluation

Avoidance functions are probabilistic in nature, so standards and internal acceptance criteria typically require repeated trials, for example 20 to 100 runs per scenario, with a defined minimum success rate. Failures are analyzed to determine whether they originate from sensor blind zones, algorithmic limitations, mechanical response delay, or environmental interference.

Applications

Obstacle avoidance function test reports are used across the entire product lifecycle and across multiple market access pathways:

• Product safety certification: demonstrating compliance with safety standards for robots and electrical appliances as part of certification marks required in target markets.
• Regulatory market access: supporting customs clearance, national product registration, and compliance documentation for markets that mandate conformity of robotic products.
• Design validation: providing engineering feedback during research and development to refine sensor placement, coverage zones, speed profiles, and avoidance algorithms.
• Quality control: establishing repeatable acceptance criteria for production line sampling and firmware release verification.
• Tender and procurement documentation: supplying objective evidence of safety performance for enterprise customers, logistics operators, hospitals, and government projects.
• Liability and risk mitigation: maintaining documented test records that demonstrate due diligence in verifying safety functions, which is valuable in incident investigation and product liability contexts.

Typical product categories benefiting from this testing include warehouse AMRs and AGVs, robot vacuum cleaners and mops, commercial cleaning robots, last-mile delivery robots, hotel and restaurant service robots, robotic lawn mowers, inspection patrol robots, and educational or companion robots.

FAQ

What standards are most commonly applied to obstacle avoidance testing?

The applicable standard depends on the robot type. Driverless industrial trucks and AMRs are generally evaluated against ISO 3691-4. Personal care robots are covered by ISO 13482, industrial robots by ISO 10218, and household cleaning robots by IEC 60335-2-2. Performance benchmarks may additionally reference the ISO 18646 series, and market-specific requirements may invoke GB, UL, or EN standards.

What test obstacles are used during testing?

Standardized cylindrical test bodies of various diameters with defined surface finishes are used to represent human legs, standing persons, and general obstacles. Additional test objects include glass panels, mirrors, black low-reflectance plates, thin cables, plush objects, and stair-edge drop platforms, depending on the product's intended environment.

What is considered a passing result?

Passing criteria vary by standard and product class. In general, the robot must detect the specified test bodies and either stop before contact or limit contact velocity and force below the allowable thresholds defined by the standard. For performance-oriented evaluations, a minimum avoidance success rate over repeated trials, such as 95 percent or higher, is commonly required by the agreed test plan.

How many samples are needed for testing?

For functional safety verification, one fully representative production-intent sample with final firmware is usually sufficient. If the test program also covers electrical safety, electromagnetic compatibility, or environmental endurance, additional samples may be required to cover destructive test items.

How long does obstacle avoidance testing take?

A focused functional evaluation covering static, dynamic, environmental, and edge-case scenarios typically requires one to two weeks, depending on the number of scenarios, repetitions, and any failure-driven retesting. Combined certification programs that include safety and electromagnetic compatibility will extend the overall timeline.

Can testing be performed during the development stage?

Yes. Pre-compliance testing during development is strongly recommended. Early evaluation of sensor coverage, stopping distance, and difficult obstacle classes such as glass and black objects allows design corrections before formal certification, reducing overall time to market.