Robot mowers and dogs: poop, safety sensors, and schedules
If you have a dog, two questions decide whether a robot mower is a good idea: will it hit the dog, and will it run over what the dog left behind. The first has published specs behind it. The second does not, and the honest answer is that no manufacturer we track publishes a pet-waste avoidance spec on any datasheet. Below is what each maker does publish for the models we publish pages for, quoted as its own claim, plus the one first-party statement we found on the poop question. We have not hardware-tested any of these machines and nothing here is a safety assurance.
Figures in this guide come from the manufacturer datasheets behind our model pages, verified 2026-08-06.
What each maker publishes about detection
Obstacle avoidance is one of the better-documented specs in this category, because it is a selling point. Every active model we publish pages for names a detection system and most name a count of recognized obstacle types. Ecovacs publishes AIVI 3D on the GOAT O1000 RTK, an AI camera paired with 3D time-of-flight sensing, rated by Ecovacs for over 200 obstacle types including small animals. Segway publishes a 300 degree RGB plus time-of-flight object detection system on the Navimow X350, with a claim of detecting over 200 common yard objects including pets, furniture and suspended items, and VisionFence with a 140 degree RGB camera on the Navimow i110N, rated for 150 plus obstacle types.
Mammotion publishes dual-camera AI vision with 300 plus obstacle recognition on the LUBA 3 AWD 3000, adding smart anti-fall protection and a bumper, and over 300 obstacle types on the YUKA mini 2 1000H. Dreame publishes 360 degree 3D LiDAR plus binocular AI vision on the A3 AWD Pro 3500 with a 300 plus obstacle-type claim. Sunseeker publishes Vision AI plus a physical bumper on the X7, stating detection of over 200 obstacle types spanning static, dynamic and low-hanging obstacles.
Two makers take different routes. Worx publishes camera-only detection on the Landroid Vision WR208, a 140 degree wide-angle HDR camera that its product page says avoids people, pets, furniture and wildlife. Husqvarna publishes onboard radar based object detection on the Automower 420 iQ, the only radar system in the set, and the only one with no camera in its published obstacle-detection description.
- AI camera plus 3D time-of-flight: Ecovacs GOAT O1000 RTK, 200 plus types including small animals.
- RGB plus time-of-flight: Segway Navimow X350, over 200 yard objects including pets.
- Camera vision: Segway Navimow i110N (150 plus types), Worx Landroid Vision WR208 (people, pets, furniture, wildlife).
- LiDAR plus vision: Dreame A3 AWD Pro 3500 and Mammotion YUKA mini 2 1000H and LUBA 3 AWD 3000, all 300 plus types.
- Vision plus bumper: Sunseeker X7, over 200 types including low-hanging obstacles.
- Radar: Husqvarna Automower 420 iQ, onboard radar based object detection.
Nobody publishes a poop spec, and that is the finding
We looked for a published pet-waste avoidance claim on every manufacturer page and datasheet behind the models we publish pages for, and found none. Not one maker lists pet waste as a recognized obstacle type, gives a minimum detectable size, or states any performance figure for it. The obstacle counts above are counts of categories a maker says its system recognizes; none of those published category lists includes dog waste.
The one first-party statement we did find is Segway's, on its own Navimow Academy blog rather than on any spec page. Segway writes that common robot lawn mowers cannot detect dog poop, treating it as a small obstacle they ignore, and then makes a conditional claim about its newer machines: that a deposit roughly 1 cm in diameter and 3 cm above the grass surface will be detected and rerouted around, while noting it depends on size and may not work for small breeds. Segway names the Navimow X4 and the Navimow i2 AWD in that passage. Neither is a model we publish pages for, and the claim is a blog claim, not a datasheet spec.
Segway's own conclusion in that article is the practical one: scoop before mowing. We are quoting the manufacturer, not adding a recommendation of our own. What the datasheets support is narrower still, that pet waste is simply not a documented obstacle category for any machine on our roster, and a buyer should plan around that absence rather than around an obstacle count.
Vision, LiDAR and radar are not interchangeable claims
The sensor family matters because the three approaches are documented differently. Camera-only systems, meaning the Worx Landroid Vision WR208, publish recognition by category with no distance or geometry spec alongside. LiDAR-plus-vision systems, meaning the Dreame A3 AWD Pro 3500 and the two Mammotion machines, add a ranging sensor and publish higher obstacle-type counts, and Dreame separately publishes a crossing height (obstacles up to 2.2 in, footnoted at 2.16 in from Dreame lab testing) which is the closest thing in the set to a published geometry figure. Ecovacs publishes both a barrier-crossing height of 3 cm (about 1.2 in) and a narrow-passage width of 2.3 ft on the GOAT O1000 RTK, the latter flagged as an Ecovacs laboratory experimental result.
Husqvarna's radar on the 420 iQ is the outlier and worth naming as such. Radar detects presence and range rather than classifying what it sees, so Husqvarna publishes no obstacle-type count at all for the 420 iQ. That is not a worse spec than a 300-type claim; it is a different kind of claim, and the two cannot be ranked against each other from the datasheets.
None of these systems is bump-only. All nine active models lead with a sensor, and three carry a bumper behind it: the LUBA 3 AWD 3000, the Sunseeker X7, and the discontinued LUBA 2 AWD 3000HX, which published UltraSense AI vision plus ultrasonic radar plus a physical bumper. If you want to compare two philosophies directly, the Navimow i110N versus GOAT O1000 RTK and Navimow i110N versus Landroid Vision WR208 comparisons set camera, RTK and time-of-flight approaches side by side.
The safety sensors makers do and do not publish
Beyond obstacle avoidance, a few models publish sensors that matter around a moving dog, and the coverage is uneven. Mammotion publishes smart anti-fall protection on the LUBA 3 AWD 3000 and lift and tilt sensors on the discontinued LUBA 2 AWD 3000HX. Husqvarna publishes lift and tilt sensors on the 420 iQ, listed under its anti-theft feature set. Dreame publishes a lift-up alert on the A3 AWD Pro 3500. EcoFlow published a lift sensor and rain sensor on the discontinued Blade.
What is not published, per model, across this set: a blade stop time on lift or tilt, a blade stop distance on obstacle detection, a minimum detectable object height, any injury or animal-safety test result, and any statement about behavior toward a moving animal as opposed to a static one. Robot mowers in the US market are built against blade-safety standards, and EcoFlow's manual is the only document in our set that names them (EN 50636-2-107 and ANSI/OPEI 60335-2-107, cited there for noise measurement). We are not going to convert a standards reference into a per-model safety spec, so those fields stay empty.
That gap is the point rather than an oversight. A dog owner reading obstacle counts can easily believe a safety case has been made. What has actually been published is a recognition claim from the maker's own testing, with the conditions unstated in most cases. Treat it as such. Our companion guide on what yard robot makers do not publish collects the wider pattern.
- Published: anti-fall (LUBA 3 AWD 3000), lift and tilt (Automower 420 iQ, LUBA 2 AWD 3000HX), lift-up alert (A3 AWD Pro 3500).
- Not published for any model on our roster: blade stop time, blade stop distance, minimum detectable object height, animal-safety test data.
- Not published for any model on our roster: pet-waste detection of any kind.
Scheduling, framed as user practice
The obvious workaround, running the mower when the dog is inside, is a household practice and we present it that way. No manufacturer we track publishes a recommended schedule for homes with pets, a pet-aware mode, or any guidance tying mowing windows to animals. Anyone telling you a maker recommends it is reading marketing copy as documentation.
What makers do publish is the machinery that makes such a practice workable. All nine active models publish multi-zone management, and all publish app scheduling: Mammotion publishes a maximum of 30 zones on the LUBA 3 AWD 3000 and 10 on the YUKA mini 2 1000H, Sunseeker publishes no zone limit on its brand specs page while its store page says up to 100, and Segway advertises multi-zone management on the X350 without publishing a maximum count. Husqvarna publishes stay-out zone support through EPOS on the 420 iQ, and Ecovacs publishes no-go areas and map editing in the ECOVACS HOME app for the GOAT O1000 RTK.
Those are the published levers: a schedule you set, and zones you exclude. Whether you use them to keep the mower and the dog apart, or to fence off the corner of the yard the dog uses, is your call and your responsibility. We are describing features, not prescribing a safety routine.
Reading the set with a dog in the yard
Nothing in the published record supports ranking these machines on pet safety, so here is what it does support. If you want the largest published obstacle-recognition claim, three models publish 300 plus types: the LUBA 3 AWD 3000, the YUKA mini 2 1000H, and the A3 AWD Pro 3500. If you want an explicit animal or pet mention in the published claim, the GOAT O1000 RTK names small animals, the Navimow X350 names pets, and the Landroid Vision WR208 names pets and wildlife. If you would rather buy a system that does not depend on classifying what it sees, the Automower 420 iQ publishes radar based detection and no obstacle-type count.
On the poop question there is no spec to shop on, for any of them. Segway's own blog says scoop it first, and that is the only manufacturer position we located. Everything else you may read about robot mowers and dog waste comes from forums and video, which is exactly the evidence standard this site exists to avoid.
If you are still choosing between navigation approaches more broadly, robot mowers without a perimeter wire covers what replaces the wire on each system, and how to choose a robot mower puts obstacle claims in line with the specs that carry a stated measurement basis.
Frequently asked questions
- Do robot mowers avoid dog poop?
- No manufacturer we track publishes a pet-waste avoidance spec for the models we publish pages for. Segway addresses it on its own Navimow Academy blog and says that common robot lawn mowers cannot detect dog poop, treating it as a small obstacle they ignore. Segway then makes a size-conditional claim for its newer X4 and i2 AWD, that waste roughly 1 cm in diameter and 3 cm above the grass will be detected, and adds that it depends on the size of the deposit. Those two models are not on our roster, and Segway's own advice in the same article is to scoop before mowing.
- Which robot mowers publish pet detection?
- Three of the nine active models we publish pages for name animals or pets in their published obstacle claims: the Ecovacs GOAT O1000 RTK (200 plus obstacle types including small animals), the Segway Navimow X350 (over 200 common yard objects including pets), and the Worx Landroid Vision WR208 (people, pets, furniture and wildlife). The Mammotion LUBA 3 AWD 3000 and YUKA mini 2 1000H publish a higher count, 300 plus obstacle types, without naming pets specifically. Naming a category in marketing copy is a claim about recognition, not a guarantee about your dog.
- Are any robot mowers bump-only?
- None of the nine active models we publish pages for is bump-only. Every one publishes a camera, LiDAR, time-of-flight or radar based detection system as its primary method. Three add a physical bumper as a backstop behind that: the Mammotion LUBA 3 AWD 3000, the Sunseeker X7, and the discontinued Mammotion LUBA 2 AWD 3000HX.
- Should I schedule a robot mower around my dog?
- Running the mower when the dog is indoors is a user practice, not a manufacturer instruction, and we present it as such. What makers do publish is that all nine active models support multi-zone management and app scheduling, which is the mechanism that makes such a practice possible. No maker publishes a recommended schedule for households with pets.