Before we start, I want to ask you what the nine points in front of ROS mean?

Turtlebot4 is a new robot platform based on ros2.0 launched in May 2022. I collected some information and compiled it into PPT, and here I will make a "Discussion on Application Scheme" of turtle 4 (OS: gossip of previous and present lives)

Speaking of turtlebot, I have to mention ROS. Speaking of ROS, I have to talk about the robot PR2 (as shown in the right figure above)

In 2007, a group of Stanford University Students (aspiring young people) came up with an idea, "is it possible for us to make a personal service robot, help us wash, cook and tidy up the room? In short, everything we don't want to do (Xiaobian also had this fantasy), and even when we are bored, they can chat and play with us... Finally, they made it, the PR2 robot just mentioned.".

As you can see in the figure above, PR2 robot can complete a series of complex application functions, such as folding towels, playing billiards, frying eggs, helping people cut their hair (bottom right). Take folding towels as an example, which was an important research sensation in the robot circle at that time (because for the first time, a robot could complete the processing of flexible objects. Although the efficiency was low, folding five towels took 100 minutes, but it pushed the robot forward a big step at the academic level).

They know that it's not easy to complete all this. Machinery, circuit and software are involved, and they span many majors. They can't do it alone. What about working with everyone? For example, it involves a set of standard robot operating system. The software framework in PR2 robot is the prototype of ROS, so ROS was born from this personal service robot, and soon became an independent robot operating system that allows more robots to program on it.

The nine points used in ROS marking are from the tortoise shell. In 2010, ros1.0 was released. In order to popularize ROS, willow garage company launched turtlebot in 2010, which is the robot standard platform officially designated by ROS. When we open ROS, a little turtle will appear, and the turtle icon even becomes the symbol of ROS.

The name turtlebot is an official mark of ROS.

Let's take a look at this picture. Have you found a phenomenon: the shape of turtlebot4 is similar to that of generation 2 and generation 1.

[return to the turtlebotxi series]

[strong base integration performance] after the release of turnlebot3 in 2017, many users gave us feedback, saying, "what you sent me was loose parts, and I asked students to spell it for at least two and a half hours. Sometimes students accidentally connected the power line of opencr upside down, and the board burned. Is there a base that can be integrated directly for me in the future?" therefore, turnlebot4 returned to the turnlebot series, and the base integration performance is stronger.

[solve the problem of battery power supply] users also mentioned that every winter and summer vacation, the robot does not charge, and it will break down when it is reused. It's a little similar to our car's battery. If you don't care about it for more than 2 months, it's easy to break if the battery is discharged too much. The battery is a consumable and can only be bought again if it breaks down. Therefore, an intelligent charging pile (standard charging pile for both versions) is equipped as standard for turtlebot4 to solve the problem of battery power supply.

[sensor upgrade] some users said that the performance of the camera on the turnlebot3 is too poor. Many times, it needs to be equipped with Intel d435 or zed2. At this time, the computing power is not enough, and it needs to be equipped with an additional TX2, which is too troublesome. Is it possible to directly improve the performance of the camera. On the turtlebot4, we used the oak camera launched by opencv company, and the product is very good (later, PPT will have a camera function comparison table)

[interaction performance enhancement] when using the turnlebot3, we don't know the battery status / WiFi status / communication status / motor status, etc., which we need to connect the display to understand. New generation turtlebot4 standard configuration UI interactive panel

[height] the standard version of turtlebot4 has made a high-rise support design

[based on ros2.0 system]

Take the overall hardware architecture of turnlebot4 for example

[perception layer] its sensors are similar to our eyes, ears

[planning level] the master controller and the slave controller (raspberry pie 4b) are similar to our brain and heart

[control layer] then its base is similar to our feet

The red part is the difference between the two robots of the turnlebot4

There are four main points:

1. The camera models are different. The standard version is oak-d-pro, and the lite version uses oak-d-lite

2. Standard configuration UI board

3. The standard version has a high-rise frame

4. The standard version has a Bluetooth handle

For more information about cameras, interested parents can get more information at Zhihu: "OAK China"

https://blog.csdn.net/oakchina/article/details/123901027#comments_21460719

1. [multi robot system]: in the future, robots will not be independent individuals, and communication and cooperation are also required between robots. Ros2 provides a standard method and communication mechanism for the application of multi robot systems.

2. [cross platform]: different robot application scenarios will lead to different control platforms. For example, the computing power performance of autonomous vehicle is certainly much better than that of AMR robot. In order to make all robots run ros2, ros2 can run on Linux, windows, MacOS, RTOS, and even microcontrollers (MCUs) without any system, so we don't have to worry about whether our controllers can use ROS.

3. [real time]: robot motion control and many behavior strategies require robots to be real-time. For example, robots should reliably find pedestrians in front within 100ms, or stably complete kinematics and dynamics solutions within 1ms. Ros2 provides a basic guarantee for such real-time requirements.

4. [network connection]: no matter what kind of network environment, ros2 can try to ensure the integrity and security of a large amount of robot data. For example, when the WiFi signal is bad, the data should also be sent to the past as much as possible, and the data should be encrypted and decrypted in the scenario of hacker intrusion risk.

5. [commercialization]: a large number of robots have come to our lives, and there will be more and more in the future. Ros2 can not only be used in the robot research and development stage, but also be directly carried in products to the consumer market, which also poses a great challenge to the stability and robustness of ros2.

6. [project management]: robot development is a complex system engineering. The whole process of project management tools and mechanisms such as design, development, debugging, testing and deployment will also be reflected in ros2, which is more convenient for us to develop a robot.

7. To meet these needs, the design and development of ros2 is not simple. Compared with standardized products such as mobile phones, Android system can also be standardized as much as possible, but robots vary widely. How to adapt to as many robots as possible may be far more complex than developing a hand-held system or computer system.

8. ROS developers face two choices. The first is to modify and optimize the architecture of ros1, which is similar to a built house. We turn it into a blank house and renovate it. But it will certainly be subject to the pattern of the original building. In the long run, it is not the best choice. They finally chose the second option, which is to push it down and start over.

9. Therefore, ros2 is a new robot operating system. Based on the successful experience of ros1, the system architecture and software code have been redesigned and implemented. Compared with ros1, it is reflected in the following points:

In this picture, ros1 is on the left and ros2 is on the right. Please pay attention to the most obvious change between the two, that is, master.

In ros1, the role of master, the node manager in the application layer, is very important. All nodes have to listen to its command. It is similar to a company's CEO, and there is only one. If the CEO suddenly disappears, the company will be in a mess. Ros2 invites this most unstable role away. Nodes can find each other through another set of discovery - self discovery mechanism, so as to establish a stable communication connection.

The middle layer is the standard communication interface encapsulated by ROS. When we write programs, we will often deal with these communication interfaces, such as publishing the data of an image, receiving the information of a radar, and the client library will call the complex driver and communication protocol at the bottom to make our development more simple and clear.

In ros1, ROS communication relies on the underlying TCP and UDP protocols, while in ros2, the communication protocol is replaced by a more complex but more perfect DDS system.

If a large amount of data needs to be communicated in the process, both ros1 and ros2 provide communication methods based on shared memory, but the names are different.

At the bottom is the system layer, that is, which operating systems ROS can be installed on. Ros1 is mainly installed on Linux, and there are many options for ros2, including Linux, windows, MacOS and RTOS.

www.turtlebote.com

www.ros.org