INDOOR POSITIONING MARKER FOR LIDAR

Abstract

The invention relates to a lidar marker consisting of a catadioptric coated with a coating, the position of the marker being obtained from the intensity reflected by the marker.

Description

Title of the invention: INDOOR POSITIONING MARKER FOR LIDAR Technical field of the invention

[0001] The technical field of the invention is that of indoor autonomous mobile vehicles, in particular vehicles which can be deployed within infrastructures such as factories, logistics centers, hospitals or public places without requiring substantial modifications to these infrastructures. Technological background

[0002] Autonomous mobile vehicles, better known by the English acronym AIV for Autonomous Intelligent Vehicles or AMR for Autonomous Mobile Robot, are robots designed to move around warehouses, production centers, logistics centers, etc., in free navigation, without guidance elements and without piloting. These robots thus make it possible to transport goods from one point to another in the warehouse, without external intervention or supervision.

[0003] AIVs use maps of their environment to locate themselves by comparing them to the measurements of their sensors. Although other technologies exist, the one used by almost all AIVs is lidar technology combined with rodometry. The robot estimates its short-term movement using rodometry (measuring the rotation of the wheels) and corrects the accumulated errors (drift) by comparing the measurements made by its lidar sensor to the map of the environment.

[0004] This solution has several drawbacks.

[0005] First, the accuracy of this solution depends on many factors, including the map resolution. Map resolution has a huge impact on the time taken for location and navigation calculations; the smaller the resolution, the longer it takes. A common, correct value is 5cm per pixel. Roughly speaking, this value also limits the location accuracy to 5cm. Such accuracy does not allow for fine-grained navigation operations such as automatic loading, also known as "docking."

[0006] Second, the location on the map needs a starting point to work. This initial position is used for the first position estimate which is then refined as the vehicle moves and corrected by comparison with the map.

[0007] For docking, the robot needs a precise local reference point and not a global one. Indeed, the important thing is not to position the robot precisely on the map, but precisely on its dock so that it can recharge or perform its task correctly.

[0008] Several solutions already exist to obtain the relative position between a robot and its charging station. There have been active infrared systems that function like a headlight to guide the robot, this requires a receiver on the vehicle and an LED diode on the charging station. We also know systems that use a camera to detect a visible marker (QR code type), this requires a camera and is often sensitive to ambient lighting (glare or night). Another known solution is the use of a remarkable shape on the charging station such as a V-shaped hollow. This particular shape is detectable thanks to lidar and allows a relative position measurement.

[0009] The use of a remarkable shape does not allow the identification of the station in front of which the robot is located. Objective of the invention

[0010] The objective of the invention is to enable precise relative positioning of the robot in order to execute a docking maneuver or a localization using a marker visible by a lidar and possibly capable of transmitting identification information in addition to the positional information. Statement of the invention

[0011] To do this, the invention relates to a marker for lidar, which is a passive and simple solution to implement. We used a catadioptric reflective strip of fixed size on which we printed a strip opaque to the wavelength range of the lidar allowing the marker to be identified.

[0012] Today, the vast majority of lidars return intensity information in addition to distance information. We used this intensity information to precisely detect the catadioptric reflector in the environment. Indeed, the catadioptric reflects much more light than standard objects, so the intensity information allows it to be clearly identified among the rest of the measurements.

[0013] However, the light intensity returned by the reflector is sensitive to the angle of incidence, which reduces the reliability of detection of the marker. To solve this problem, we have identified a coating, which can be dependent on the wavelength of the lidar, to be added to the reflector acting as a filter to overcome this problem. This transparent filter makes it possible to smooth the returned light intensity and solves the sensitivity problem.

[0014] The size of the marker allows its position and angle relative to the robot to be precisely calculated. This calculation is done at the sensor's measuring frequency. In addition, this allows the servo algorithm to perform a docking maneuver with an accuracy of the order of +-5mm.

[0015] For the marker identification code, we made the calculations using the specifications of the lidar used, we determined the following pattern. The marker consists of one or more bands absorbing the wavelength of the lidar and positioned on a catadioptric reflector to optimize the contrast, the catadioptric and provided with a specific coating to treat the angle of incidence. The position and width of this opaque band define the marker identifier.

[0016] This will create a fixed-size "Black-white-black-white-black" pattern that is easy to detect in the set of measurements returned by the sensor.

[0017] The essential characteristic of this solution is the use of a catadioptric element to be able to use the intensity returned by the lidar sensor in order to precisely locate this element.

[0018] The use of a filter to smooth the measurement is secondary, but makes the measurement reliable, regardless of the angle of incidence of the ray. This is what makes the method reliable enough to be usable.

[0019] Using more than one band for the identifier allows the number of possible identifiers to be increased.

[0020] The marker can be used without an information encoding strip. In this case the marker has no identifier, but can still be used to perform a specific maneuver.

[0021] The use of a strip allows for a strong distinction. The reflector and the strip facilitate the identification process.

[0022] Using a single strip simplifies manufacturing and identification, but reduces the number of usable identifiers.

Claims

Claims

0.

Images