Anyone who knows a little about robot vacuum cleaner knows that it has two types: random and planning, which is very understandable; the random sweeping robot has nothing regular when sweeping the floor, and only starts to change the comprehensive path after colliding with obstacles, so when you watch it work, you should not be anxious, seeing the garbage there, it just does not pass; and it is a sharp contrast is the planning, In contrast, the Robot vacuum cleaner is not as blind as a random sweeper, and it has several things in mind.
Now, more and more sweeping robots are planning and planning navigation; they are still quite different from each other.
This requires SLAM (Simultaneous Localization and Mapping), a localization and map building technology, which can be implemented using either LiDAR technology, vision technology, or inertial navigation.
Inertial navigation relies on gyroscopes, accelerometers and other inertial sensors to obtain a position, velocity and additional information. It is widely used in defence fields such as aircraft, missiles, ships, submarines, tanks, etc. However, as the cost decreases and demand grows, it is gradually expanding to commercial fields such as floor sweeping robots.
Affected by the accuracy of inertial navigation devices, the inertial navigation process has errors. Over time, errors will continue to accumulate in a large area of the complex ground environment; the disadvantages of inertial navigation will gradually emerge, can not be suitable for planning-type cleaning tasks.
SLAM based on LIDAR relies on a laser ranging sensor to scan the room; when the laser is projected onto the obstacle, it will form a light spot, and the image sensor will calculate the distance to the centre of the laser ranging sensor according to the pixel number of the soft spot, and then combine with the robot's algorithm to build a room map and position the cleaning in real-time.
Although both can draw a map, the final map drawn has apparent differences.
The map drawn by the inertial navigation robot is more elementary and straightforward, without detailed trajectory route, and even only at the end of the home can barely see the home layout map.
In addition, some comprehensive technology that does not meet the requirements of the machine is also equipped with a laser sensor. However, the actual use of its path planning will find that it still relies on inertial navigation, so the final map drawn out of the accuracy is not high, and the map was drawn above using an inertial navigation machine and does not give full play to the advantages of laser navigation technology.
Visual navigation is also called vSLAM, there will be a camera on the sweeping robot, similar to laser SLAM, calculate the obstacle distance through the collected point cloud data, then based on the monocular, fisheye camera vSLAM scheme, using multi-frame images to estimate its positional change, and then calculate the distance from the object by accumulating the positional change, and perform positioning and map construction.
In addition, even if the same positioning navigation technology is used, the algorithm technology directly impacts the sweeping results. The last thing I want to say is that sweeping robot is still a relatively young industry. Technology accumulation still needs time, so you need to improve the overall algorithm capability through a firmware upgrade constantly.
Perhaps you are interested in the following news article?
An smart home clean robot that can simultaneously sweep and mop the floor and automatically wash the mop cloth is reliable; talk about my personal experience.