Skip to main content

Natural Navigation Automated Guided Vehicles

Step by step “free” or natural feature navigation is entering in the AGV (or Autonomous Mobile Robots, AMR) industry. 

   

The concept is simple, “free” navigation, so a navigation without need of hardware installation such wires, tapes, reflectors. 

   

In this article, I will try to explain you how do Automated Guided Vehicles with free navigation work and some interesting concepts like the SLAM Navigation concept?

  

Discovering the Natural Feature Navigation

 

Autonomous Mobile Robot with Natural Navigation

    

What is Natural Navigation for AGVs? 

       

In general, Automated Guided Vehicles with Natural Navigation are able to navigate autonomously by identifying and mapping the surrounding area. Natural Navigation is the most evolved agv navigation method.

   

The driverless robots with natural navigation rather than following a magnetic tape on the floor (or perform reflector laser triangulation) they can, for example, identify a defined wall and navigate at a given distance from it.

 

 

In fact, there are several technologies under the “natural navigation” umbrella.

   

The most typical technology is based on contour recognition (contour-based).  The vehicles with this technology navigate by identifying and mapping the surrounding area with different types of sensors used by mobile robots.

      

It is important to understand that the type of self-driving robots performing natural navigation are called Autonomous Mobile Robots (AMR) rather than Automated Guided Vehicles. 

    

By the way, do not miss my article explaining the differences between agvs and amrs. Otherwise you can register and download this whitepaper full of useful info.  

     

AGV vs AMR Whitepaper

     

Click here to download the whitepaper

   

Ok, let's go ahead... 

    

Table of Contents

    

What is SLAM? What is slam technique?

SLAM Navigation in the AGV Industry

What does simultaneous localization and mapping SLAM software do?

How does AGV robot map its environment?

But how does AGV know where “it” is?

What are AGV Natural navigation advantages?

What are free Navigation disadvantages?

When is useful to choose AGVs with natural Navigation?

 

     

   

What is SLAM based technology? What is SLAM Navigation technique?

    

The term SLAM is the acronym for Simultaneous Localization And Mapping. So a SLAM based AGV is be able to map the environment and localize itself in that map. 

  

Do not worry, I'll explain how do robots with slam navigation work. 

    

  

   

SLAM based navigation is step by step entering in the AGV industry. In fact, I'm convinced that the SLAM techonology will, sooner or later, substitute other traditional navigations methods such magnetic tape navigation.

        

Basically, the mobile robot maps the environment with different types of sensors such as LiDARs or Vision Cameras, and it is able to localize itself in that map. 

 

The robot creates a "theorical" map and then compares it with the "real map" that it "sees" while running. 

       

It is a great solution but the environment must count on well defined specifications.

       

There are good suppliers of AGV Navigation Software developing and marketing Natural Navigation Technologies.  An AGV Manufacturer could buy from a Navigation Technology supplier or could develop its own Navigation Technology.  

    

There are two key points to overcome before we’ll see SLAM everywhere:

     

  • Free navigation could be complicated in chaotic and variable environments. The robot must be able to "see" a good percentage of the environment otherwise it does not know where it is. 

         

  • Even if there are some free or low-cost SLAM software, a good SLAM navigation software is more expensive that other traditional navigation methods. 

      

In these TABs you can find some of the most important of AGV Navigation Software  suppliers for the Mobile Robot Industry... do not be shy and contact them for further details.

    

 

    

   

What does simultaneous localization and mapping SLAM software do?

    

SLAM is mainly composed by:

   

  • The first thing to do is to “map” the environment to create a “theorical map” that is stored in the robot or management system “brain”.

 

  • Once the robot is working, it performs “Landmark Extraction”. The robot navigates and maps the “real” environment or contour.

 

  • Data association. The data received while navigating is compared with the “theorical” environment previously recorded in the first stage.

  

  • State estimation. The self-driving robot estimates where it is based on the “real data” received compared to the “theorical data” stored. This info is completed with other data arriving from other IMU (Inertial Measurement Units) such encoders, etc.

  

  • State update and landmark update. State and map is recalculated considering old and new data.

 

All these calculations are given by complex algorithms that can vary depending on the developer. Moreover, results depend on the tools and instrumentation used for data acquisition.

  

For this reason, SLAM is more a “concept” than a single algorithm or method.

 

   

How does the SLAM driverless robot map its environment?

     

There are several methods and sensors to map or track the environment and estimate the mobile robot positioning. Each  manufacturer uses one or a combination of the following ones.

    

LiDAR sensors 

     Lidar Sensor for mobile robot navigation

 

LiDAR sensors are widely used in the mobile robot industry. The LiDAR scanner measures the distance to a target by illuminating the target with laser light and measuring the reflected light with a sensor.

AGV with Lidar Sensor Mapping the environment

  

Differences in laser return timing and wavelengths can then be used to make digital 2-D or 3-D representations of the target.

 

The term LiDAR is the acronym of Light Detection and Ranging

 
 

The LiDAR emits laser beams while running and the coming back info is used to map the environment and identify the AGV position.

 

There are several types of LiDAR sensors, scanning lidars, solid state lidars, etc. 

  

If you wish to know more about LiDAR sensors and technology, you cannot miss this recorded webinar:  Latest Trends in Lidars for Mobile Robots applications by SICK

  

Click here to watch the Webinar

   

  

Sonar

  

AGVs can also use sonars for navigation. Sonars are old fashion compared with LiDAR technology but are less expensive. Their measurements are worse than those of laser scanners.

  

Vision Guided Navigation for Vision Guided Vehicles (VGVs)

   

Vision Guided Vehicles often use stereo or triclops systems to measure distances. Using vision resembles the way humans look at the world and thus may be more intuitively appealing than laser or sonar. Also there is a lot more information in a picture compared to laser and sonar.

  

The main problem is that vision technology provides huge amount of valued information that is difficult to process needing complex and advanced algorithms. Of course, technology and processing capacity is improving so Vision Guided technology is becoming easier, reliable and affordable.

      

But how does the mobile robot know where “it” is?

     

There are three main strategies for “matching” information.

  

First strategy: Scan-to-scan

  

The data received from the sensors (for example the LiDAR) is used to estimate the position of the AGV between two consecutive scans.

  

As a result, we have updated and accumulated the positioning of the vehicle.

  

This method does not depend on any predefined map, so it is very useful when map is missing or to create an initial map.

  

The main problem is that error increases over time without possibility to correct it.

  

Second Strategy: Scan-to-map matching

   

In scan-to-map matching, data is used to estimate the position of the vehicle by matching measurements directly done on field with a theoretical map previously stored in the Management System or in the Robot's brain. 

    

This strategy does not accumulate errors on time but could create many problems in case the scanned environment does not match with the theorical map. For example, in chaotic production environments.

    

Third Strategy: Combination with odometry

   

None of these strategies is completely satisfactory and reliable, so here comes the third strategy: Combination of scan-to-scan, scan-to-map and robot's inertial and odometry data to overcome the problems and to provide accurate and reliable performance in real-world settings and applications. 

   

An important aspect of the SLAM is the combination with the odometry data (acceleration, wheel encoders, gyroscope, etc). The goal of the odometry data is to provide an approximate position of the robot by asking to its internal odometry sensors.

   

The slf-driving robot mixes all the information given by the measured environment, the theorical environment and its own sensors to provide as much as possible accurate positioning.

  

  

   

What are the advantages of Natural Navigation for AGVs?

  • It’s the future. This techonology presents many advantages and it is becoming more reliable and convenient. 

   

  • Easy to install. Installation is relatively fast, just turn-on your robot, map the enviroment or draw a line in your PC and the robot will be able to navigate work immediately.

   

  • Low installation costs. Installation is fast so you will need few time to install the system. It means lower workforce cost. 

      

  • No invasive installation. Natural Navigation AGVs do not require any kind of external infrastructure or civil work.

      

  • No costs due to implementation maintenance. There is not any infrastructure to maintain (like magnetic tape), so no cost. 

What are the disadvantages of Natural Navigation for AGVs?

  • Mobile Robots with Natural Navigation could have localization  problems in chaotic workplaces due to variable environment.

     

  • Good and reliable natural navigation is still more expensive than other traditional navigation methods.

     

  • Lower accurate positioning compared with other navigation methods.

       

When is useful to choose AGVs with natural Navigation?

      

As anticipated, in my opinion, the Natural Navigation will substitute other types of navigation such as magnetic navigation, optical navigation ,ecc.  The main manufacturers of AGVs and AMRs are developing and including this technology in their robots.

          

The main concern today about Natural technology is its reliability in variable environments such production lines, where there’s continuous movement of people, manned means, boxes, pallets, etc. In this conditions a mobile robot with natural navigation might not find where it is because the contours are continuosly changing.

        

For this reason, Natural Naviagtion  is a great solution for mobile robots  when you have well defined profiles and environments with fixed structures such walls, columns.  

    

Natural Navigation is the best solution for applications like: 

    

 

 

 

  

In general, in any environment with low level of “confusion”.

       

If you want to know in-depth technical details about SLAM technology, check this link.

    

Related articles...to  learn more about Natural Navigation

 

AGV Navigation Software Suppliers with Videos, who's innovating in the navigation industry? 

    

AGV Navigation Methods : Types, Comparison, Pros and Cons - Illustrated Guide 

     

 AGV vs AMR  What should you buy? Differences? Pros & Cons? - Winner?

    

AMR Applications: 7 Key Applications that AGVs can't perform


Linkedin Written by Alfredo Pastor Tella (agvnetwork editor).

Follow me on LinkedIn... let's create a mobile robot community to discuss and learn about these outstanding systems.

 
Category: AGV TECHNOLOGY