Electronic Luggage Follower

Electronic Luggage Follower César Nuñez, Alberto García, Raimundo Onetto, Daniel Alonzo, Sabri Tosunoglu Department of Mechanical Engineering Florida International University 10555 West Flagler Street Miami, Florida 33174 [email protected], [email protected], [email protected]

ABSTRACT In this paper, we describe the concept of building a robot able to pursue a specific person through an airport while assisting with carrying that person’s luggage. After a review of the current devices available for performing these tasks, we described our approach that aims to develop a platform that could send and receive a signal that would provide a simple and practical means for the robot to determine a path and follow it that does not require the use of internal maps and the ability to self-localize. In particular, the approach is based on a control system able to execute obstacle avoidance and target following behavior. Also, a relative location device based on a signal emitter (placed on the target person) and a directional sensor (placed on the mobile platform).

Figure 1. Conceptual Design

1. INTRODUCTION In this research, we consider how robots can act in concert with human behavior. Our aim is to develop a robot capable of maneuvering through busy airports behind its owner while hauling his or her luggage. In this paper, we consider the technique to be used to trail people and report on the realization of a mobile robot capable of following a person. In order to follow a human, a mobile robot needs to know the position of the person and must be able to determine its own path in order to follow his target. We consider a

method using an infrared light-emitting device and a receiver. In order to prevent collision with obstacles, ultrasonic sensors are used to detect objects that may be in its path. We present the effectiveness of our approaches by showing the experimental result using a real mobile platform. In this article, we describe the research carried out in the attempt to develop a human-following mobile robot. It was decided to provide the followed person with a transmitter installed on the back heel of their shoe that would broadcast infrared signals visible to the robot’s receiver that would allow it to detect the relative direction of the person. The robot should also be able to exhibit an effective obstacle avoidance behavior and to integrate obstacle avoidance with target following and exploration behaviors. Along with using infrared receivers and transmitters, the robot emit pulses of ultrasonic waves to avoid any obstacle that may come between itself and its owner. In the design of the platform that would carry the luggage, many factors must be taken into consideration. It must be light, but of a strong enough material to hold the average luggage. Compatible motors and microprocessors must be selected in order to make all the components work seamlessly. Through the use of modern day technology, mobile robots have become autonomous enough to carry out numerous applications. We explore this research and review the ways robots can provide services following commands without any physical contact with people. Our goal is to develop an autonomous robot that could follow a person with the purpose of assisting them in the daunting tasks of dragging heavy luggage across long distances in airports. While various tasks could be applied to such a robot, i.e. guidance, information supplying or escorting, in this paper, we consider all the components and devices required to complete the task of a creating follower robot with the capabilities of avoiding obstacles while carrying luggage.

2. PROJECT CONCEPT The Electronic Luggage Follower (ELF) will be a luggage that will follow the user throughout any flat surface without the need of the user to use force to drag it. No effort will be applied by the user in order to carry different load magnitudes.

Florida Conference on Recent Advances in Robotics, FCRAR 2010 - Jacksonville, Florida, May 20-21, 2010

1

3. OBJECTIVES • • • • • •

A robot easy to be used and to be manageable by any person. A wireless system made of a transmitter part and a receiver part connected to the luggage. An optional way to carry the luggage in case of any problem. A sound structure and base design to resist load, different temperatures, and external forces. A luggage with an attractive and innovative exterior design. A security system that the user can be free of worries of his or her luggage being stolen or left behind.

4. FORM AND FUNCTIONALITY The product comes in the same sizes as any standard luggage. The system inside the bags is resistant to any load or external force. The luggage functions with several different sensors, ultrasonic sensor and touch sensor. These sensors prevent the bag from bumping onto obstacles, such as people, wall, and other things on the floor. The form of the upper part of the luggage is made round to allow aerodynamics. The based form follows the upper part producing a sense of surface smoothness. Inside the base, all the components are arranged on the back part in a stable and practical way. The functionality of the luggage as said before is to follow a person in either a manual or a wireless way. Ultrasonic sensors are in charge of the wireless function of the luggage. Touch sensors are used to manually drag the luggage.

The biggest challenge in this system is to control the forward and stopping motion. As the user pulls the rope a little, the sensor must be activated and the luggage must go forward.

6. ANALYSIS TO SET UP THE INITIAL CONDITION 6.1. Velocity Analysis: To perform gear analysis of the system, desired velocity of the luggage, input torque of the system is needed. The desired velocity of the luggage is 6.56 ft/s. This is the measured velocity for a person walking at a fast pace. The input torque of the system is obtained from the design specifications from the motor. By calculating the desired velocity, v, of the system, the angular velocity of the wheel can be obtained, using the following equation.

/

(1)

Where ω is the angular velocity and r is the radius of the wheel. After performing calculations, the angular velocity of the wheel is approximately, 546.7 rpm. Using this value, the gear train is analyzed. Gear 3 is attached to the same shaft as the wheel, so the angular velocity of the gear is the same as the wheel. Using the following equation, the angular velocity of G2, ω2, can also be found, as well as the angular velocity of gear 1, ω1.

(2)

5. CHALLENGES Dragging a heavy luggage causes a strain on the person’s body and causes discomfort. The product being developed will end the discomfort of having to drag luggage all around any flat surface. It will also eliminate luggage being left behind, or stolen. The original concept was developed by New York architect Peter Yaedon. However, his design has not been developed and does not yet meet airline requirements. The challenges of ELF can be classified in two areas; the wireless design and the easily drag design. For the wireless system, the challenge is to achieve the physical separation of an ultrasonic sensor. Ultrasonic sensors are sensors where the transmitter and receiver work together as a one piece. Originally the transmitter sends a signal that bounces in any surface and then is received by the receiver. The lap of time in recorded and the distance is estimated. Separate this sensor is our biggest challenge. Once the physical separation is achieved, another challenge appears. The sensor should estimate the distance according to the magnitude of the feculence instead of the time that takes the signal to leave and return to the sensor. For the easy to drag system, the challenge is based on the accuracy of the manual command. The user will use a rope to “drag” the luggage. By using touch sensors, the rope will be an activator for motion. As the rope is moved to either left or right, the luggage will rapidly move to the corresponding side so that no force is required by the user.

Where n is the speed of the gear, and N, is the number of teeth of that gear, a is a representation of what gear speed is needed, and b is the gear with the known speed. So the values of the angular velocities, in rpm, for all three gears are as follows:

n1 = 109.33 n 2 = 182.20 n3 = 546.70 The motors used in the prototype provide a maximum speed of 100 rpm at free rotation, so the desired velocity cannot be accomplished. With the motor loaded with the gears and adding small weight on the system, an input velocity of 50 rpm is assumed. Using equation 2, the actual values for the system used are obtained, and they are as follows (rpm):

n1 = 50

n 2 = 83.33 n3 = 250 With these angular velocities, the output velocity of the system is approximately 5.99 ft/s. These values prove that the motor used can provide enough velocity to move the system at about 6 ft/s. However, it is very important, when


2

working with rational mov ving parts to cheeck that there iss m, to carry a larrge load and stilll enough torqque on the system travel at the desired speed.

6.2. Torq que/Force An nalysis: In order to get g the luggage to start moving, a great amountt of torque neeeds to be provided. To start off ff, the coefficientt of friction between b the flo oor and the wheeels needs to bee determined. For the calcculations of thiis luggage, thee n between rubbber and concretee coefficient of static friction was found. The coefficient of static frictionn is used becausee i at rest. The vaalue of the coeffficient of frictionn the system is is 1.0. Once this value is chosen, then thhe friction forcee between thee wheels and thee ground can bee found with thee following eqquation:

Ff = μFn

luggaage, and everydday use, a higgher torque mootor is needeed.

7. R RECEIVER R AND TRA ANSMITER R S SET UP In ordder to use a wirreless system ass desired, two seeparate circuiits must be devveloped. One ciircuit will proviide the ultrassonic sensor traansmitter with the power andd input signal required (see Figure F 2). The other o circuit willl adapt the siignal received so s that it can be b used by the MicroContrroller.

(3))

Where Ff is i the friction force, µ is thee coefficient off friction, andd Fn is the normaal force. The fricction force is thee force needed to get the lugg gage to start mooving. This forcee can be pluugged into the Power equatioon. The powerr equation is a function of wo ork and time.

P = Work W / Time

(4))

However, thhe work required d by the wheel is unknown, butt it can be dettermined by usin ng equation 5.

W = Forcce x Distance

(5))

Substitutingg equation 5 into o equation 4, thee power equationn becomes

P = (Forcce x Distance)//Time = Forcee x Velocity (6))

Figure 2. 2 Transmitter Circuit In Fiigure 2, the 9V V battery that provides p energy to the frequency generator and the ultrasoonic transmitter at the same time is visualizzed. A power coonditioner is plaaced so that the t voltage can be reduced to 5V 5 before enteriing the frequency transmitterr (the frequency transmitter usedd works at 5v)). The frequencyy transmitter willl create a sine wave w of 40 KH Hz that will be sent s by the ultrassonic transmitterr.

Now, all thhe values needed to calculate the power aree known, andd the power requ uired by the whheels is 229.6 ft-lbs, carryingg a maximum lo oad of 70 lbs. Using U this power, or the correesponding torquee, along with thhe gear analysis, the torque required r by the motor is determ mined. The gearr attached to the t wheel, G3, reequires a torque of about 230 ft-lbs. This torrque is transmittted to G2 througgh force Ft32 andd this force is calculated using g the following equation: e

F32t = (2T)/d

(7)) Figure 3. 3 Transmitterr Circuit 2

In this case d is the pitch diiameter of the gear, and T is thee torque transsmitted by this gear. Using thhis equation, thee torque transsmitted through each e gear can bee obtained.

Taable 1. Calcula ated Torque Values V T1 T2 T3

51.58 5 ft-lbs 139.9 1 ft-lbs 229.6 2 ft-lbs

o T1 is the torqu ue required to move m the luggagee The value of at the desirred speed of 6..56 ft/s. The motors m purchasedd cannot provvide such torquee so for commeercial use of thee

Once the signal is recorded by the ultrasonic sensor receivver, a signal ampplifier, based on 741 Op Amps, is used in ordder to increase the signal freqquency. After thhat, the signal is used by thee PIC16F690. Here H the analog signal cominng in is transforrmed to digital in order to quantify it later. A programminng code is used to convert the digital frequency of the inpuut signal into a range r number between 0 andd 99 (as the freqquency increasess, the number beecomes biggeer). Then, a pulsse with modulaation is created so that this number n can be transformed innto the percenttage of elapseed time for a duuty cycle that will w be created att 50Hz.

Florida Confference on Rece ent Advances in Robotics, R FCRAR R 2010 - Jackso onville, Florida, May M 20-21, 2010 0

3

This duty cycle is totally compatible with the microcontroller, so from there on the conditions of the program can be written without further problems.

8. HUMAN DECTECTION USING ULTRASOUND SENSORS First, we equip the person with a sensor-emitting device and make the robot detect this device using an ultrasound sensor. In order to determine the distance to the human, we use the sending and receiving time of the signal between the sensors. The person carries this device perpendicular to the ground. By receiving the signal of this device, the robot is able to determine the distance to the human due to the time interval between the two signals. It can also determine the direction taken by the device by gathering information based on the ultrasound sensors time interval of the signal sent by the human carried device. If we consider how a mobile robot could estimate the position of a specific object by using ultrasonic sensors, it is also possible to find the distance to the object by using the pulse echo method. The use of a sensor consisting of several receivers enables us to determine where the obstacles are and avoid them. However, it is not possible with this method to recognize a specific object among several others. We considered that it would be possible to differentiate a specific object if the ultrasonic wave emitted by the robot performing measurements can be separated from the ones emitted by a transponder carried by the object whose position is to be detected. The robot transmits an ultrasonic pulse A and the object which has an transponder transmits another ultrasonic pulse B after detecting pulse A while the pulse A is returned by the reflection at an obstacle.

9. MICROCONTROLLER SELECTION It is important to properly select the correct microcontroller based on the objectives set out. Based on the design goals, the major components that will influence the performance of the electronic luggage follower are identified as follows: • Million of Intructions per Second (MIS) • Number of Programming Pins • Power Consumption • Programming Language

9.1. PIC16F877A The PIC16F877A is an 8-bit programmable integrated circuit made by Microchip. This PIC has forty pins, with various different purposes. For example, there are PORTB and PORTC which are bidirectional I/O port and can be software programmed. A major advantage of this chip is the versatility of all the 40 pins, since the peripherals are spread out over the pins. Furthermore, it maximizes the number of external devices to attach when compared to other microcontrollers. Another feature of this chip that is favorable for the user is each pin has only two to three functions, making it easier for the user to decide the purpose of each pin. Additionally, it has a wide operating

voltage range of about 2.0 V - 5.5 V, and an internal (4 MHz) or external clock select. Finally, the chip has a cpu speed of 5 MIPs. A main component of the PIC16F877A that will be used toward the performance testing of the microcontroller is the universal synchronous/asynchronous receiver/transmitter (USART). This component is what gives the microcontroller the ability to translate and send data to the computer. This device changes internal information to serial data so that it can be sent on a communication line. Finally, this chip is very inexpensive ($4.75) and can be easily programmed using MikroC.

Figure 4. Microcontroller PIC16F877A 9.2. 25AA640 The 25AA640 is a small 64kbit Serial Electronically Erasable PROM (EEPROM). EEPROM is a chip that can be programmed multiple times, and it does not need to be taken out of the computer to program it again. A major advantage of this chip is that the information that is saved on it does not erase when the chip is turn off. This chip can be bought for a dollar, and can be programmed also using mikroC. Contrary to the PIC16F877A, this chip only has 6 pins. Two of the 6 pins are programmable, one as the input and one as the output. For the electronic luggage follower, only two programmable pins is a major disadvantage since more than one sensor will be installed as input and more than one sensor will be installed as output. Additionally, this chip has a maximum clock frequency of 1 MHz, which compared to other PICs is very slow. Finally, a major disadvantage towards the study of the performance of this chip is that it does not contains UART, and the information cannot be sent back to the computer to read it.

Figure 5. Microcontroller 25AA640


4

9.3. DSPIC30F4011 The DSPIC30F4011A is a 16-bit digital signal programmable integrated circuit made by Microchip. The advantage of this DSPIC is that it is a combination 16-bit microcontroller and a DSP. The 16-bit microcontroller will enhance the performance while the DSP will implement a high computation speed. Similar to the PIC16F877A, it has 40 pins that can be used for programming external devices, analog-to-digital converter, interrupts, EEPROM memory, UART, etc. A major advantage over the other two options is the internal speed of the microcontroller. This chip has a 7.37 MHz internal clock and other components that will enhance the cpu speed to 30 MIPs. Also, if the user is trying to obtain information about the external devices connected to it, this microcontroller has two different UARTs, and can be an advantage when trying to test any devices. Furthermore, this chip can be found in the market for about $5.00 and it can be easily programmed just as any other microchip product using mikroC.

Figure 6. Microcontroller DSPIC30F4011

send a signal for 5 µs to the ultrasonic sensor. After the signal has been sent, the microcontroller then waits for 1 µs to change from output to the input mode. When the pin is set to input, a while loop is initialized to wait at input until the signal is received. Finally when the chip obtains the signal, it will send 5 Volts to the speaker. The speaker provides an indication that the sonar sensor is detecting an object.

Figure 7. Signal Propagation After analyzing the specifications of the chips, the microcontroller that will be best suitable for the electronic luggage follower was chosen. The decision was made over several factors, such as price, number of pins, CPU speed, and programming language. Based on the datasheet of all options, the PIC16F690 was chosen as the optimal microcontroller for the design. Although it was the smallest, it proved to have the sufficient amount of pins for the platform. Even though the DSPIC30F4011 was the fastest, it uses a different programming language that is more complicated, and it will lead to an over-designed platform, as the speed exceeds the requirement. While the PIC16F690 had a cost of $1.91, the others cost nearly 5 times more.

9.4. PIC16F690 The PIC16F690 was the last programming chip that was analyzed made by Microchip. This is an easy-to-use chip that has 20 pins, with various different purposes. The microcontroller is small in size which acts as a major advantage towards the purpose of the project. Furthermore, it is an 8 bit chip with nano-Watt technology for low power battery consumption. Finally, the PIC16F690 was purchased with a kit which facilitated the development of computer programming. The PIC16F690 also has an 8MHz internal clock and has a CPU speed of 5 MIPs. A disadvantage of the microcontroller is that it does not contain the UART feature. Without this component, transforming information from the chip to the computer was not possible. Finally, this chip is very inexpensive ($2.00) and can be easily programmed using MikroBasic.

10. TESTING AND RESULTS The algorithm in Figure 8 is a representation of the programming of the sensor. The program performs in such a way that a pin is first initialized as output and used to

Figure 8. Programming Algorithm After the selection of the microcontroller, a test was conducted in order to ensure the performance was accpetable. A circuit board was set-up in a way that the chip was connected to both an ultrasound sensor and a speaker. Afterwards, the set-up was completed by developing a test program. This program instructs the microcontroller to send a signal to the ultrasonic sensor and then when it receives the signal, it activates the speaker. This program proved the high-performance and the simplicity of the microcontroller’s features.


5

Figure 11. Thermal Analysis on the Base of the ELF (Aluminum Alloy 2014) Figure 9. Testing Sensors

11. HUMAN FOLLOWING BEHAVIOR In order to prevent collision with obstacles, the robot should track the route followed by its target. The robot calculates the distance between the actual position of the human and the previous location of the human. If the distance the human moved is greater than a value decided in advance, the mobile platform will increase its speed until it reach the right time interval between the sensors signals. In order to estimate its position, the vehicle uses odometer and computes the location of the human relatively to the position of the vehicle. The location of the human is recorded by the robot in a global referential. The robot can adjust its speed according to the number of points recorded in the list that exist between the actual pose of the robot and the location of the human. By doing so, the robot can track the path followed by its target.

The second material used was the Plastic ABS and this plastic by applying the heat was deformed in some little zones. As it can be seen in the next figure, the maximum deformation is produced where the motors are; in other words, the deformation is maximum where heat is situated. The plastic ABS has passed the thermal test, yielding a very similar result comparing to the Aluminum Alloys 2014. The plastic ABS passed the thermal test.

12. STUDY CASES 12.1. Thermal Analysis: The thermal analysis was performed on the base. The reason is that the base is the part most exposed to heat. For this analysis, 25°C is used as room temperature and a heat power of 50W is also used. The team performed this thermal test with different types of materials, some failing while others passed the benchmark. The materials that passed the thermal test were selected as viable option to be used as the final material for the electronic luggage follower platform. This was important because a priority was luggage being safe enough to be in any regular temperature. The materials used to do this test were aluminums and plastics. The following figure shows how the aluminum alloy 2014 past the thermal test. The part covered by the blue color is the part that was not affected by the heat produced by the motors, and as hotter the piece as redness the color. This indicates on the figure that the base is safe as far as heat concerns. The most heated parts are where the wheels will be connected and the rest of the structure is maintained in a low temperature. That problem can be fixed by changing the piece affected and by adding some more material in the critical zone.

Figure 10. Thermal Analysis on the Base of the ELF (Plastic ABS) 12.2. Force Analysis: The force analysis is made on the base and the shelves since these are the parts that will receive or resist the higher loads. The base will support the entire weight of luggage which, we estimate to be approximately 50 lb. The shelves will contain clothes, shoes, etc. so they should be able to support more than 5 lb each. For the force analysis on the base, 70 lb force is applied. The load was exaggerated in order to have a stronger luggage that can sometimes carry extra weight. The force analysis is divided in three parts: deformation, displacement, and stress. The force analysis will be expressed in three different figures. Aluminum alloy 2014 is the first material to be evaluated for the force test on the base of the ELF. The following figures show that the aluminum alloy 2014 force test failed by deformation in the middle part of the luggage. This deformation is produced basically because of the lack of support in the middle region. The solution that the team suggests for this deformation is to add to our design an all axis movement wheel that will not affect the movement of the luggage and provide the support needed to prevent


6

failure in the critical zone. Figures 12 displays the force analysis results for the aluminum alloy 2014.

Figure 14. ELF Prototype

Figure 12. Force Analysis Stress on the Base of the ELF (Aluminum Alloy 2014) The second material evaluated is the plastic ABS. The plastic ABS performed quite well compared to the aluminum alloy 21014. The deformation showed in the next figure is very little and because the force applied is greater than the regular force used by the luggage users, the material can be considered as one of the final options. The base using the plastic ABS presents some critical deformation in the front part of the luggage. Figure 13 shows FEA results for plastic ABS.

We presented a method to achieve human following behavior as a first step toward the development of an intelligent escort robot moving along with a person. We showed two methods to realize the human following behavior using ultrasonic sensors. As future work, further investigation is needed to assess the robustness of these methods and develop new approaches to cope with situations when the robot cannot detect humans. The team is also planning to make experiments using sound generation to inform the status of the robot to the human in order to realize a smoother interaction. Furthermore, the team is developing an accompanying behavior model for the human. Once the stage of building a clever robot capable of moving along with the humans is achieved, we will then model the relation between human and intelligent robot interaction.

14. References [1] Bianco R., Caretti M. Nolfi S. Developing a robot able to follow a human target in a domestic environment. In A. Cesta (Ed.), Proceeding of the First Robocare Workshop. Institute of Cognitive Sciences and Technologies, CNR. Roma: Italy, 2002.

Figure 13. Force Analysis Stress on the Base of the ELF (Plastic ABS)

13. CONCLUSION In this paper, we addressed various design stages of a follower robot that is envisioned to carry the owner’s luggage through crowded air terminals and follow the owner. The machine is developed to be affordable for a new invention that can lead to new applications to aid humans further.

[2] Krohn A., Beigl M., Hazas M., Gellersen H., Schmidt A. (2005). Using Fine-Grained Infrared Positioning to Support the Surface-Based Activities of Mobile Users. Proceedings of the 5th International Workshop on Smart Appliances and Wearable Computing (IWSAWC), Columbus, USA, 2005. [3] R. Bischoff, .Advances in the Development of the Humanoid Service Robot HERMES,.Field and Service Robotics Conference , pp.156-161, 1999. [4] Y. Hayashibara, Y. Sonoda, T. Takubo, H. Arai and K. Tanie, .Localization and Obstacle Detection for a Robot for Carrying Food Trays, IEEE/RSJ International Conference on Intelligent Robots and Systems, pp.695-700, 1999.


7