networked sensors for robot monitoring and control



Journal of the Technical University – Sofia Plovdiv branch, Bulgaria “Fundamental Sciences and Applications” Vol. 21, 2015 International Conference on Engineering, Technologies and Systems TECHSYS 2015 BULGARIA

NETWORKED SENSORS FOR ROBOT MONITORING AND CONTROL ROBERT KAZALA, ALBENA TANEVA, STOICHO PENKOV Abstract: In this work a network development for monitoring and robot controlling is presented. Solutions related to the network protocols and standards for predefined application are summarized. The main goal is focused on net operation by using low energy protocols. In the developed project for sensors data exchange was use MQTT protocol. The real test and verification with sending and receiving data between the connected nodes are made. The advantages of the networked sensors are given. Key words: Networked system, IoT Ethernet, MQTT protocol

1. Introduction Monitoring of air and land pollution on a large area, of radiation during nuclear power plant failure are complex and sometimes dangerous tasks [4]. In order to determine an accurate map of the degree of pollution is necessary to perform multiple measurements at different locations. This task over a long period of time is performed. It is case of highly dynamic phenomena, such as spread of contamination or irradiated particles in the air, means that the results do not reflect the real value. A solution of such problem is to use mobile nodes equipped with relevant networked sensors for data collection. Several robots can be organized and formed a grid. This way can be covered large areas and different surfaces. Can be used mobiles and flying robots. One of the main factors for such system development is the implementation cost. In order to achieve it is appropriate to use small, lowpower mobile nodes, which are controlled and supervised by a main robot. Mobile robotics platform that grants access to robotic resources over a network, more specifically, the robot’s communication model within the platform is introduced in [1]. The paper presents a WSN extension to provide communication between the robot and the platform server over IEEE 802.15.4, instead of IEEE 802.11 (WiFi). This paper is focused on configuration and development of the networked sensors for robots monitoring and control. A number of works review shows variety of protocols and communication networks [2, 3, 4, 5]. Generally efforts are focused

on problems with network delays, packet dropouts, address channel limitations related to the packetrates, but there is no universal solution. In the paper problems definition, development suggestions and maintenance for networked sensors are made. This work summarizes solutions related to the network protocols and standards for predefined application. Proposition of mobile robot based sensor network with specified low energy protocols are also presented. 2. System design and configuration It is a difficult task to gather information from many and different points. In a stage of problem definition the task is related to a network design and configuration of robots equipped with sensors. The possible solution is to have a grid of networked robot nodes. If there is a lot of equipment, it is important to find a low-cost solution with low energy consumption. In order to reduce costs it is necessary to minimize the tasks of networked robots (to develop a grid of nodes). This network allows reducing the requirements for computational power and can eliminate the data processing of the separate node. Whereas in the network design have to be include few different nodes (or robots). Therefore cable connections are not relevant or suitable. In such case of many distributed robots cooperation the option is to use radio communication. The solution performs the data transmission from the sensors to the central computer (or robot). This allows using of low-cost microcontrollers. Hence, they must be equipped with relevant communication interface. Another

Copyright  2015 by Technical University - Sofia, Plovdiv branch, Bulgaria

ISSN 1310 - 8271

cost factor influencing for the system is communication interface. The power consumption, easy operation and reconfiguration of the system are also given and discussed in [6]. Nowadays there are many radio communication standards, but not all are well suited to the task. The ability and widely used popular, not expensive modules with easy connection to the Internet is very important. One solution that comes to mind is the standard IEEE 802.11 (WiFi). It allows connecting multiple devices to the network, and easy integration with the Internet. However, it has the disadvantage of relatively high demand for energy, which is particularly important in the case of using small flying robots to collect data. Among the standards for devices with low power the most interesting solution is IEEE 802.15.4. IEEE standards are for the physical layer of the OSI model. They allow transferring data. But do not provide a convenient way to control communications and meeting the requirements of QoS (Quality of Service). Accordingly, the next step is the selection of a transport layer protocol. The choice of transport protocol is not an easy task. If there is need to connect to the internet there are basically only two options TCP and UDP. These protocols allow multiple devices to communicate effectively using the Internet. However, they require determining ways, for different types of data, how they are divided and stored in frames. In the case of a system design for collecting data in order to reduce the workload, related to the organization of data exchange, it is possible to use application-layer protocols. 3. Application-layer protocols Depending on the solution and hardware layer there are different and possible choice options. One of them is to use industrial automation protocols. The arising problem in this case in general is high cost of implementation. Another solution that can be implemented for data exchange is the use of open Internet protocols, e.q.: • Hypertext Transfer Protocol (HTTP) • Representational State Transfer (REST) • XMLHttpRequest (XHR) • Comet • WebSocket These protocols are used to exchange data between clients, which are usually web browsers and servers providing http websites. They allow transferring data in a standardized way and providing easy integration of a robot with the Internet. However, they have the disadvantage of a large overhead of transmitting data frames, based on the complexity and data structures used by these

protocols. This limits their use to efficiently transfer data between the small robot nodes. Today, one can still consider the option of using protocols developed for the exchange of data between devices connected to the Internet, the socalled M2M or devices called IoT (Internet of Things). Among these protocols can be distinguished:  Constrained Application Protocol (CoAP),  MQTT,  ZeroMQ,  Advanced Message Queuing Protocol (AMQP),  Streaming Text Oriented Messaging Protocol (STOMP). By analyzing network protocols, extra attention should be pay to the model they use for data exchange. Many of the technologies used in modern computer systems use a data exchange model referred to as Request-Response. In these systems, the client sends a request to the server and then waits for a response. In the case of websites this kind of communication in most cases is working correctly. However, when you try to use such a model for data exchange in the sensor network you can encounter some difficulties. For the clients receiving the data it is necessary to continuously polling the server, therefore, communication is carried out even if the data has not changed. In the robot sensor network, it is better to use a mechanism in which the robots send the sensors data to the main computer, when they change or in the specified time interval. Such a solution is possible using the Publish-Subscribe method. In this method, the data publishing modules send it to a server called the broker, which then sends the data to clients subscribed to certain information. Using these methods of data exchange allows the clients to receive not all the information sent by the node, but only the data that interest. There is also no need to constantly calling the modules that generate information about the data. 4. MQTT protocol Among the protocols previously described one of the most appropriate is MQTT protocol (Message Queue Telemetry Transport), details in [6]. It was designed in 1999 for transferring data from telemetry devices. The main goal of the designers was to create an efficient protocol to transfer data from devices with limited hardware resources, which is equipped with a lowperformance microprocessors and a small amount of memory. Also expected to work in networks with severely limited bandwidth for data transmission. The protocol uses a publish-subscribe method and transmits the data over TCP/IP or UDP.

In its implementation requires a special computer called a messages broker. The task of the broker is to collect messages and sending them to devices interested in specific information. Figure 1 shows the organization diagram for data exchanging between Publishers and Subscribers by MQTT Broker.

MQTT Broker

Subscriber 1

Publisher 1 Topic 1

Subscriber 2 Topic 2

Publisher 2 Subscriber 3

Fig. 1. Organization of data exchange in MQTT. MQTT protocol messages are assigned to names that are topics. In context of the client and the broker, there is no need to configure the topic. The client sends a message to a specific topic. If there is a particular topic the broker will update its data, in the absence a new topic will be created automatically, to which will be assigned the information transmitted in the message. Topics may be organized in a hierarchical manner using the separator in the form of a forward slash (/). This allows us to organize data in a broker in a manner similar to the file system. Example topic for networked grid nodes may have the following form: network/NODE/sensor/DATA An important feature of the MQTT protocol is the ability to manage the quality of service by implementing QoS (Quality of Service). It allows you to manage the way to deliver a message and confirmation of its receipt. There are three levels:  QoS 0 - At least one delivery.  QoS 1 - At least one provide.  QoS 2 - Only one supplied. 5. Sensors network proposals for the robots monitoring and control system This work is devoted on configuration and investigations of the sensors network for monitoring and control. A solution is found out by combination between the standards 802.15.4 and Wi-Fi 802.11. These two protocols can be used on one device and allow obtaining the Gateway to the Internet. On the fig. 2 is shown the general proposal of standards for the separate and for the master nodes in the network. In the case study of low energy nodes the IEEE 802.15.4 is used with 6LoWPAN protocol (shown in the upper grid of the fig.2). For other cases with more powerful nodes it is easier to use

IEEE 802.11 Wi-Fi protocol (lower figure part). Hence as to the each node is made MQTT client installation. It can communicate with broker installed on the main robot. Advantages of the system are: easy maintenance; low cost; low energy consumption; reliability and security; easy connection to the Internet. The networked sensors system (the grid) can be enlarged with up 1000 nodes. In the particular test application for the sensors network development it is obtained the simple topology which uses only IEEE 802.11 equipped nodes. There are three connected nodes in this case study. Furthermore it will be possible and very easy to extend the node numbers if necessary. The idea is associated with networked sensors for robots to perform various tasks. Each “node” presents robot sensor and is connected with the main robot with IEEE 802.11 protocol. IEEE802.15.4  Node 12  Node 13 

Node 14 

Node  11 

Bridge  node  

Node 1

Node 4  Main  Robot 

Node 2 

IEEE802.11

Node 3 

Fig. 2. The developed and investigated sensors network In application test for the robot sensors network only IEEE 802.11 nodes were used. To build the system standard microprocessor modules equipped with wireless communication interface were used: -the main control module - mobile computer installed on a central robot, -the data collector - a prototype system Connected LaunchPad Tiva with Wi-Fi module from Texas Instruments. The most important element of the system using MQTT protocol software is a broker running on the main control computer. In the project only open source software were taken into consideration:

Copyright  2015 by Technical University - Sofia, Plovdiv branch, Bulgaria

ISSN 1310 - 8271

• Mosquitto, • Moquette, • Mosca, • RabbitMQ. Of the listed above brokers the implementations of Mosquitto was used. This is due to the support of the development of this version by the Eclipse Foundation, and high availability in different versions of Linux repositories. Once installed, the broker can be configured via a file mosquitto.conf. After installation Mosquitto requires no configuration and in most cases and applications work properly in a default configuration. To test the operation of the broker, the universal client applications can be used that are available as independent programs, web applications run in the browser or applications for mobile devices. The project used mqtt-spy application, which is an independent application written in Java. After testing the operation of the broker the next step was to implement clients sending and receiving data. The MQTT protocol, there are many versions of client libraries for various programming languages. The project used the following versions: • Eclipse Paho MQTT C/C++, • Eclipse Paho Python. For all Connected Lunchpad Tiva modules PAHO MQTT Eclipse C / C ++ library was used to write applications. PAHO Eclipse Python was used in the client installed on your computer to remotely control the robot and covering August sending the data generated from the control falls. The last set of libraries mqtt.js used to build web applications, whose task is to monitor the work of the individual modules robot. An important step in designing communication layer using MQTT protocol is to determine the structure of topics and related messages. In current system only test messages were send, which allows reading data from nodes sensors. 6. Conclusion The paper shows an example of the use of open network protocols in networked mobile robot sensor system. MQTT protocol allowed for the use of standardized methods for data exchange in the sensor network. Also greatly simplified the integration of new nodes, the use of nodes information from other systems and integrating with the Internet. The use of open protocols simplifies software development work, especially when it consists of a large number of independent nodes. Also simplifies the maintenance process, since it is possible to read information about the robot and the status of individual sensors without having to use special tools. Future work will be focused on real time robot network implementation.

REFERENCES 1.

2.

3. 4.

5.

6.

7.

Souza R., L. Agostinho, F.Teixeira, D.Rodrigues, L. Olivi, E. Guimaraes, E.Cardozo. Control of Mobile Robots Through Wireless Sensor Networks. XXIX Simpósio Brasileiro de Redes de Computadores e Sistemas Distribuídos, pp.805-818. GreenPeak Technologies (2015), Wireless Communication Standards for the Internet of Things, White paper Silicon Laboratories, Inc., The Evolution of Wireless Sensor Networks, 2015. ITU-T, Ubiquitous Sensor Networks (USN), ITU-T Technology Watch Briefing Report Series, No. 4, February 2008. Suzuki T., Kawabata K., Hada Y. and Tobe Y., Deployment of Wireless Sensor Network Using Mobile Robots to Construct an Intelligent Environment in a Multi-Robot Sensor Network, Advances in Service Robotics, Ho Seok Ahn (Ed.), ISBN: 978-953-7619-02-2, InTech 2008. X. Wang, G. Xing, Y. Zhang, C. Lu, R. Pless, C. Gill, Integrated Coverage and Connectivity Configuration for Energy Conservation in Sensor Networks, ACM Transactions on Sensor Networks, Vol. 1, No. 1, August 2005. David Johnson Tim Stack Russ Fish, Flickingery D. M., Stoller L., Ricci R., Lepreau J., Mobile Emulab: A Robotic Wireless and Sensor Network Testbed, IEEE INFOCOM, April 2006 ACKNOWLEGEMENTS

The authors would like to acknowledge Ministry of Education and Science of Bulgaria, Research Fund Project No: DFNI-I02/6/2014. Department of Industrial Electrical Engineering and Automatics Kielce University of Technology Al. Tysiaclecia Panstwa Polskiego 7 25-314 Kielce, Poland Department of Control Systems Technical University–Sofia, Branch Plovdiv 25 Tsanko Diustabanov St. 4000 Plovdiv Phone (optional):+ 359 32 659 585 E-mail: [email protected] E-mail: [email protected] E-mail: [email protected]