Comparison · 6 min read
Navigation determines whether a robot vacuum covers your rooms systematically, finds its dock, and respects the areas you mark off. Many mapping models use lidar, cameras, or both; simpler robots may navigate with bump sensors, wheel tracking, and other sensors without saving a room map. The key distinction is between knowing where to clean and detecting what is lying in the way. A robot can map your home well and still swallow a charging cable.
ⓘ Quick answer
Lidar is a strong starting point for mapping without room lighting. For cables, toys, or pet waste, look for a dedicated obstacle-avoidance system with good independent test results. A camera used for navigation does not automatically recognise those hazards. Models combining lidar with camera or depth-based obstacle detection can address both needs, but sensor count alone does not guarantee better performance or justify a higher price.
How lidar works
Lidar uses laser light to measure distances to surrounding surfaces. A common design scans around the robot from a rotating sensor, collecting measurements that software combines with movement data to build a map and estimate the robot's position. Because the sensor supplies its own light, it does not need a lit room to measure distance.
Mapping and obstacle detection are different jobs. A conventional horizontal lidar scan measures surfaces at the sensor's height. A chair leg may intersect that scan, while a cable lying on the floor may sit below it. That is why a robot with excellent lidar mapping can still need separate sensors to avoid small objects.
Strengths
Lidar provides distance measurements in bright rooms and complete darkness. Combined with capable software, it supports systematic coverage, room maps, and reliable positioning. Room selection and no-go zones are useful features to check, but they depend on the robot's app and mapping software; they are not exclusive to lidar.
Weaknesses
A fixed turret adds height and can prevent access beneath furniture. Mirrors and glass can also complicate readings by reflecting or transmitting the laser light. Basic mapping lidar measures geometry without necessarily identifying a cable, sock, or pet accident. Some models use embedded, solid-state, or retractable lidar, so a raised turret is not unavoidable.
How camera navigation works
Camera-based navigation often uses visual simultaneous localisation and mapping, or vSLAM. Software follows visual features across successive images and combines them with movement information to estimate the robot's position and map its surroundings. Performance depends on camera placement, lighting, visible features, and the software interpreting them.
Camera navigation and object recognition are separate capabilities. A camera can help the robot locate itself without identifying floor hazards. Recognising cables or shoes requires suitable views of the floor and dedicated recognition software. Some robots combine those functions; others pair lidar mapping with a camera, depth sensors, or structured light for obstacle detection. Structured light projects a pattern and measures how it changes across nearby surfaces.
Strengths
With appropriate software, cameras can distinguish object types and help the robot decide what to avoid. Camera-based mapping can also support systematic routes, saved maps, and room-specific cleaning. Designs without a raised sensor turret may fit beneath lower furniture, although the robot's actual height matters more than its navigation label.
Weaknesses
Visible-light cameras need enough illumination and useful visual detail. Darkness, glare, or low contrast can reduce performance. A lamp or compatible infrared system may help, but does not guarantee reliable navigation and hazard recognition in every condition. Cameras also introduce image-handling questions. Neither lower prices nor poorer map accuracy can be assumed simply because a robot uses a camera.
Choosing between them
Match the robot's tested behaviour and app features to your home:
- You want it to clean at night or in dark rooms. Lidar can map without room lighting. Check low-light obstacle avoidance separately: a robot may navigate successfully while its camera struggles to recognise objects.
- Your floors have cables, toys, or a pet. Prioritise dedicated obstacle avoidance and tests using similar hazards on similar flooring. Recognition varies by object size, contrast, and placement. Independent testing shows that some robots still struggle with small objects and pet waste. Pick up loose cords and check for pet accidents before scheduled runs.
- You have a large or multi-room home. Lidar is a useful starting point, but also check saved maps, room editing, no-go zones, and recharge-and-resume support. These features determine whether the robot can finish the job with little intervention. Camera-based models can offer them too.
- You need it to fit under a low sofa or bed. Measure the narrowest clearance and compare it with the robot's stated operating height. Allow space for rugs and uneven flooring. Both turret-free camera designs and low-profile lidar designs are options; check how retractable sensors behave beneath furniture.
- You are uncomfortable with a camera indoors. Choose a model explicitly described as camera-free. A lidar robot may still have an obstacle-recognition camera. Even camera-free models can upload maps, cleaning history, or other account data, so check the privacy policy and which features require cloud access.
💡 The question to ask a seller
Ask: 'What data leaves the robot, for which features, and can I disable those transfers?' On-device recognition describes where an algorithm runs; it does not prove that images remain on the robot. Obstacle-photo review, remote viewing, diagnostics, and product-improvement settings may involve separate data transfers. Ask about image and map uploads, retention, deletion, and the effect of opting out. For example, iRobot's privacy policy describes opt-in image collection, illustrating why individual settings matter alongside the sensor specification.
Frequently asked questions
Is lidar safe for eyes and pets?
Check the exact model's manual for its laser classification. Many robot vacuums use Class 1 laser systems under IEC 60825-1, a classification intended to provide safety during normal use. For example, the EZVIZ RS20 Pro manual identifies its radar and line laser as Class 1. Follow the manufacturer's instructions, keep the sensor housing intact, and do not dismantle or modify the laser assembly.
Can a camera robot clean in the dark?
It depends on what the camera does and which other sensors the robot uses. Visible-light camera navigation generally needs illumination. A suitable light or infrared-capable system can help, while a hybrid robot may navigate using lidar. However, maintaining a route and recognising a dark cable are different tasks. Look for evidence that both navigation and obstacle avoidance work with the lights off.
Do combined systems justify the price?
They can if you need dependable mapping and frequently leave small objects on the floor. The benefit is fewer rescues and less preparation, provided the obstacle-avoidance system works well. On mostly clear floors, lidar mapping without advanced object recognition may meet your needs. Compare actual models: the price also pays for features such as the dock and mopping system, and a premium sensor package does not guarantee better debris pickup.