Abstract
With the increasing elderly population, the information intellectualization demand also rises substantially. At the same time, intelligent information interaction system- (IIIS-) controlled wheelchair’s automatic obstacle avoidance techniques have become extremely important in an intelligent social context. Therefore, this paper proposes an IIIS to help disabled groups and some older people enjoy the barrier-free information design and improve product information manipulation experience. Firstly, the functional requirements of the IIIS are analyzed, and the overall scheme of the IIIS is designed. Then, the control mode of IIIS is determined. Secondly, following research on the obstacle avoidance algorithm (OAA) based on fuzzy control, the obstacle avoidance control strategy is formulated for the elderly accessible wheelchair IIIS. Finally, experiments are designed to verify the control performance and simulate the automatic OAA of the designed IIIS. The experimental results indicate that the touch screen control system (TSCS) of mobile phone (MPHN) is more sensitive for the elderly accessible wheelchair control information system. By comparison, the joystick control system (JCS) is more straightforward to manipulate than the touch screen. The practicability of these two control methods is very strong. These two independent control methods meet the control performance requirements and improve the automatic obstacle avoidance performance of IIIS. The proposed IIIS based on TSCS and JSC can respond to input commands in real time, help the IIIS avoid obstacles, and acquire images, among other functions. Therefore, the proposed intelligent wheelchair design improves the interactive performance of the intelligent operating system, facilitates the travel of the elderly and the disabled, and has good practical value.
1. Introduction
Today, the overall world population is aging. For the first time in history, older people might outnumber children in years to come. Like many developed countries, China has also entered an aging society. Nevertheless, Chinese society has some unique characteristics. For example, as the Chinese tradition extends, respect and care for the elderly are deemed a virtue, and people pay more attention to the life of the elderly. Meanwhile, with the rapid social development and family planning policies, most Chinese families have undergone substantial structural changes. As a result, many elderlies are left unattended or forced to live alone, aggravated by rapid urbanization that has attracted most young people out of their hometowns. Not to mention the many disabled who still demand further social attention to lead a better life. These social matters have become a concern for both the state and society. Thus, more attention is being paid to improving the life quality of the elderly and the disabled. Notably, the present work holds that barrier-free information communication is the most important for these particular groups. It has high practical significance under science and technological development and the intelligent era [1].
Information interaction design is an interdisciplinary subject of information design, interaction design, and sensory design. Information design deals with the organization and representation of data: transforming data into valuable and meaningful information. Interaction design is carried out according to the sensory needs of different users to improve information efficiency. In this paper, the interaction design based on fuzzy control theory is carried out for the elderly accessible wheelchair operating system. It was aimed at improving the user experience of elderly wheelchair users. At present, wheelchairs have become a scientific and reliable resort for many people with physical impairment, including some elderly and physically disabled groups. Mainly, wheelchairs can either be controlled manually or electronically (automatically). Apparently, the manual operation requires family or professional nurses’ assistance, sometimes not feasible for low-income or solidary groups. On the other hand, the automatic wheelchair control system performs poorly in obstacle avoidance, especially in multiobstacle environments. The handling of information interaction systems needs to be improved. In a nutshell, these wheelchairs are far from intelligent [2, 3]. For example, most available wheelchairs are not equipped with an antilock braking system (ABS) to avoid skidding [4]. Accordingly, Wu et al. [5] introduced a wheelchair-intelligent ABS structure to improve user safety, including the adaptive neural fuzzy inference system (ANFIS) and friction coefficient estimation system. The friction coefficient estimation system used a particle filter (PF) to quickly adapt to nonlinear states and unknown environments. The system provided accurate ABS braking control according to the road’s friction coefficient range. The ABS used a gyroscope to detect wheelchair acceleration and angle and then calculated the parameters. The user could click the on-chair stop button to activate the ABS. The results implied that the proposed ABS could effectively shorten the braking time and distance and improve wheelchair riding safety. Luo et al. [6] designed a toilet-assisted wheelchair system to solve the problems of difficulty getting up and unsanitary toilet treatment for the disabled, the elderly, and other wheelchair users. The designed wheelchair system adopted a single-chip microcomputer (SCM) for intelligent information control. It had two main functions: assisting wheelchair users in getting into the wheelchair and collecting the pollutants discharged. Experiments showed that the toilet-assisted wheelchair was more ergonomic in helping users get up and more convenient and hygienic in treating toilet waste than similar products.
With the rapid development of mobile terminal platform technology and the increasing popularity of the smartphone (SPHN), more and more services have been offloaded to terminals. Android operating system has been widely used as the mainstream mobile operating system because of its free features and powerful functions. Employing the Android platform to realize more daily applications, especially mobile portable medical devices, has become an important research field. Arguably, with the growing trend of worldwide population aging, the intelligent wheelchair will get a higher market share shortly. Meanwhile, all nations invest more resources into intelligent and more user-friendly wheelchair Research and Development (R&D). Moreover, designers need to fully consider the actual scenarios and users’ practical needs during the marketization of intelligent wheelchairs. For example, there is an increasing number of solitary elderly wheelchair users. Monitoring functions and more accessible human-computer interaction (HCI) design might be of great importance for these groups. Based on the above analysis, the present work designs a wheelchair intelligent information interaction system (IIIS) based on obstacle avoidance algorithm (OAA).Then, the wheelchair IIIS’s obstacle avoidance function is improved using fuzzy control theory. The obstacle avoidance performance is verified through experimental design. Therefore, the proposed wheelchair IIIS can improve the convenience of system operation, facilitate the elderly and the disabled to move more conveniently and quickly, and has certain practical value
2. Materials and Methods
2.1. IIIS Information Scheme Design of Wheelchair Operating System
2.1.1. Information Function Requirements of Wheelchair Operating System
A wheelchair IIIS is designed to facilitate the movement of the elderly and some particular groups. Safety is the top priority. Meanwhile, the wheelchair should be designed with easy-to-accessible and detailed information operating systems. Overall, the design should adhere to safety, reliability, and accessibility principles [7, 8]. The functions of the wheelchair IIIS read as follows: (1)The intelligent wheelchair can turn around, change direction, and emergency brake during driving. These systems should be made simple and with remote monitoring functions(2)Upon encountering obstacles, the intelligent wheelchair should avoid obstacles automatically(3)To enable users to control better, the intelligent wheelchair should be able to monitor the surrounding environment in real time(4)On irregular road sections, the wheelchair should not fall, and it can automatically alarm against an emergency
2.1.2. Overall IIIS Scheme Design
Generally, the intelligent wheelchair comprises a mechanical structure and operating system. The operating system is similar to human organs and is the core part of the entire intelligent wheelchair, whereas the mechanical structure is identical to the human body, carrying the hardware circuit of the operating system [9, 10]. There are two most common wheelchairs: traditional wheelchairs and simple electric wheelchairs. Traditional wheelchairs generally need extra assistance to manipulate and, thus, increase the labor cost. The electric wheelchair controls the movement through the joystick, with certain potential safety concerns in use [11, 12]. Against these problems, this section designs a wheelchair IIIS with a personnel monitoring function and automatically obstacle avoidance function.
The intelligent wheelchair mounts a chassis driven by a four-wheel robot. The seat is installed on the moving mechanism. Three ultrasonic sensors are installed in the forward direction of the wheelchair. The control joystick and control button are installed on the right armrest of the wheelchair, which can control the moving direction and accelerate and decelerate, brake, and whistle. The wheelchair IIIS equips with both the touch screen control system (TSCS) and joystick control system (JCS). Additionally, a camera is set above the wheelchair to collect the environmental information. Then, the wireless communication module transmits the camera images to the SPHN so that the user can understand the surroundings in real time, to better control the wheelchair. Figure 1 presents the overall framework of the designed wheelchair IIIS.
Figure 1 reveals that the microcontroller (MC) is the core part of the wheelchair IIIS, responsible for controlling the wheelchair operating system. The motor drive module controls the forward and the reverse rotations, acceleration, and deceleration. Meanwhile, the motor drive module detects obstacles and transmits the information to the MC to avoid the obstacles automatically. Wireless fidelity (Wi-Fi) module is connected with the MC through the serial port to realize the remote control. The angle sensor module measures the inclination angle of the wheelchair during movement to judge the wheelchair’s real-time states. In front of the steps, the wheelchair IIIS can timely brake to prevent the wheelchair from falling. Even if the wheelchair falls accidentally, the monitoring alarm module will send an alarm signal in time to the guardians of the user in the form of short messages. Simultaneously, the wheelchair IIIS can alarm to remind the surrounding people to help the vulnerable users. The camera module collects the environmental images around the wheelchair. It transmits them to the SPHN for users to observe the situation behind the wheelchair and operate safely.
2.1.3. Design of Intelligent Wheelchair Control Mode
Most wheelchair users are either the elderly or with impaired legs and feet. Therefore, the advantages and disadvantages of control methods for different groups differ significantly, which should be a top concern in wheelchair IIIS design. So far, two control methods prevail in the market: TSCS and JCS. Figure 2 illustrates the existing form of HCI mode.
(1) Voice Control System. The user sends instructions to the operating system in the form of speech. Upon receiving user instructions, information identification and response are activated. Thanks to the fast-growing speech recognition technology in recent years, voice control is seeing broader applications in intelligent terminals. However, voice control has some defects. For instance, unclear voice commands can hardly be correctly recognized. Thus, the current voice control technology heavily depends on users’ voice clarity. The interactive effect of voice information needs to be improved.
(2) TSCS. Users can simply slide their fingers on the SPHN screen to send commands to the operating system. It is widely used because TSCS is easy to learn, convenient, and simple to operate. Besides, it can also be remotely operated through SPHN.
(3) Visual Control System. This control method collects the user’s gestures and other limb movements through the visual sensor. The advantages are easy to operate and highly reliable. The disadvantages are high requirements for hardware equipment, complex control algorithm, cross-disciplinary methodology, high price, and some technical problems. Therefore, the visual control mode will not be considered in the proposed wheelchair IIIS.
(4) JCS. Mainly, the JCS sees applications in electric wheelchairs. It is relatively straightforward and flexible, so it is widely used.
Therefore, the proposed wheelchair IIIS combines the TSCS with JCS to simultaneously realize traditional control and remote control based on the analysis above.
2.1.4. Design of Obstacle Avoidance Mode of the Wheelchair IIIS
Wheelchair IIIS uses sensors to collect environmental information to avoid obstacles automatically. In recent years, there have been many kinds of sensors. Still, only three types are often used in robot obstacle avoidance, namely, ultrasonic measurement, infrared measurement, and image acquisition [13].
(1) Ultrasonic Measurement. The so-called ultrasonic sensor detects objects through ultrasonic waves, as shown in Figure 3. The ultrasonic sensor module sends out the short wavelength sound signals to detect direction. The sound wave will be reflected from the object along the way, and the MC will pinpoint the object according to the calculated reflection time. Such a detection method is relatively simple and is more robust against environmental influences.
(2) Infrared Measurement. Like ultrasonic measurement, the wave (light) is reflected once it reaches an object. The reflection time is used for measurement. Nevertheless, unlike ultrasonic measurement, it is highly sensitive to environmental impact.
To sum up, the ultrasonic measurement is relatively simple and more robust to environmental factors. Thus, the proposed wheelchair IIIS employs ultrasonic sensors.
2.2. Elderly Accessible Wheelchair IIIS Design Based on Fuzzy Control
Obstacle avoidance system (OAS) is the crucial link of wheelchair IIIS, so there are high requirements for its reliability. Specifically, it should mainly concern two functional requirements: the sensor control module and OOA. Because of the complexity of the wheelchair environment, there are many uncertainties, such as motor effects, which might affect the performance of wheelchair IIIS [14]. To this end, this section applies the fuzzy control algorithm in the OAS of wheelchair IIIS.
2.2.1. Fuzzy Control
The idea of fuzzy control is to control objects with computers over people. Essentially, it is a computer simulation of people’s ideas of control [15]. Artificial control experiences will be coded into fuzzy rules. Thus, fuzzy control is an intelligent control approach, as demonstrated in Figure 4.
Figure 4 indicates that the fuzzy controller contains the fuzzy interface, inference engine, defuzzification interface, and knowledge base. The fuzzy interface maps the accurate input variable into the corresponding fuzzy quantity to promote the controller to identify the output quantity. The inference engine can change fuzzy input into fuzzy output. The defuzzification interface is responsible for transforming the unrecognizable fuzzy quantities into recognizable quantities. Lastly, the knowledge base includes a database and rule base. The database can save the data in the controller. At the same time, the rule base is a fuzzy rule collection written in fuzzy language simulating human thought [16, 17]. Fuzzy controllers work like this: firstly, it fuzzifies the output and then turns it into the fuzzy quantity the fuzzy system recognizes. Secondly, it infers the input based on fuzzy control rules through an inference engine to obtain the output of the fuzzy system. Thirdly, it transforms the output into identifiable quantities and transmits them to the controlled objects. Finally, it completes an entire control process. Further, the fuzzy controllers compare the input and output and calculate the deviation, and a new round of the fuzzy control process begins to obtain a new control quantity. Overall, the fuzzy control is an iteration process based on one such circle [18, 19].
Advantages of fuzzy control are as follows: (1) fuzzy control is a rule-based control, which directly adopts language control rules. The starting point is the control experience of field operators or the knowledge of relevant experts. It does not build an accurate mathematical model of the controlled object, so the control mechanism and strategy are easy to accept and understand. The design is simple and easy to apply. (2) It is easier to establish language control rules from the qualitative understanding of industrial processes. Thus, fuzzy control is very suitable for those objects whose mathematical model is difficult to obtain, dynamic characteristics are challenging to grasp, or changes are very significant. (3) The model-based control algorithm and system design method easily lead to significant differences due to different starting points and performance indexes. Nevertheless, a system language control rule has relative independence. The fuzzy connection between these control rules makes it easy to find a compromise choice, so the control effect is better than that of the conventional controller. (4) Fuzzy control is designed based on heuristic knowledge and language decision rules, which is conducive to simulating the process and method of manual control, enhancing the control system’s adaptability, and making with a certain level of intelligence. (5) The fuzzy control system has strong robustness and has significantly weakened the influence of disturbance and parameter change on the control effect. Hence, it is especially suitable for controlling nonlinear, time-varying, and pure delay systems.
A fuzzy controller can be one-dimensional (1D), two-dimensional (2D), or even three-dimensional (3D). The 1D fuzzy controller contains only one input, the deviation between system output and expected value. Equation (1) expresses its fuzzy relationship:
In Equation (1), represents the input subset, and denotes the output subset.
The 2D fuzzy controller has two inputs: deviation and the deviation derivative to time. At present, the 2D fuzzy controller is most widely used. Equation (2) calculates its fuzzy relationship:
In Equation (2), and represent the input subset, and indicates the output subset.
The 3D fuzzy controller has three inputs: deviation, the first derivative of deviation to time, and the second derivative of deviation to time. Because the 3D fuzzy controller has many input variables, multiple control rules are required, increasing system calculation and the control time while reducing system response. Therefore, 3D fuzzy controllers are hardly seen in practical applications.
2.2.2. Design Method of Fuzzy Controller
(1) Fuzzification. The fuzzification process collects sensor information and converts them into language quantity. Then, it establishes membership function (MF) for different language quantities and finally expresses it through a fuzzy set [20]. In general, the method of changing digital variables into fuzzy linguistic quantities reads as follows: (1)Single point set method: it fuzzifies the input data ’s exact value and then turns it into two separate fuzzy sets. The fuzzy set is marked as , and the single point set is as follows:
Importantly, the single point set method only transforms the digital quantity into fuzzy quantity on the surface, but it does not transform the digital quantity into fuzzy quantity in essence. (2)Triangular fuzzy set method: the noisy fuzzy input is a fuzzy random variable. Such MF is regarded as an isosceles triangle. In fuzzy triangular sets, the vertices are the average of random numbers, and the length of the bottom edge is ( is the standard deviation (SD) of a random number)(3)Normal distribution function method: it is the most commonly used MF among fuzzy variables, as expressed by
In Equation (4), refers to the expected value, and represents the SD.
(2) Knowledge Base. The knowledge base is at the core of the fuzzy controller and is a set of expert-generated rule bases [21]. Fuzzy rules can be established through parameter settings, such as state, output, and control variables. Meanwhile, the knowledge base should consider the compatibility and integrity between control rules. Usually, the description accuracy depends on the input of fuzzy variables. Precisely, more input leads to higher accuracy and more complex rules until the control rules become too complex to maintain proper efficiency and control effect. Therefore, the fuzzy variables should be set according to the specific situation. Each fuzzy rule determines a fuzzy relationship. Then, iteratively, the fuzzy relationship of the whole system can be obtained through the logical relationship between fuzzy statements, as shown below:
(3) Fuzzy Reasoning. Simply put, fuzzy reasoning compares the fuzzy input with the fuzzy rule base and then outputs the reasoning results according to the reasoning method. It often uses the likelihood inference method and the min-max method.
(4) Defuzzification. Generally, the control signal is an exact quantity, but fuzzy reasoning generates a fuzzy quantity, so it is necessary to fuzzify the output [22]. Barycenter, weighted average, and maximum membership methods are often used in practical applications.
2.3. Fuzzy Obstacle Avoidance Information Interactive Control Strategy for Wheelchair IIIS Operating System
The right priority selection strategy is adopted to improve the efficiency and accuracy of wheelchair OAS. In simple terms, the wheelchair will prefer to move towards the right when it gets too close to an object. Nevertheless, if the obstacle is happened to be on the right, the wheelchair will choose to move to the left. Since obstacles’ relative positions to the wheelchair are random, the wheelchair OAS might choose different paths every time. Figure 5 manifests the wheelchair OAS flow.
Further, practical applications have to consider dead corners in wheelchair operations. In that case, the wheelchair will be surrounded by obstacles from three directions. Therefore, a reverse path selection design must be included in wheelchair OAS. Specifically, against multiple obstacles, the wheelchair OAS first selects a path, remembers it, and then selects the new path based on historical memories if the path is still obstructed. The operation will repeat itself until all obstacles are avoided.
2.4. Software Implementation of OAA
Subsequently, OAS inputs the obstacle distance information and the direction angle between the wheelchair and the target and outputs the wheelchair steering angle. Here, the OAA of the wheelchair IIIS is coded by C programming language. Figure 6 demonstrates the core part of the fuzzy control-based wheelchair OAA.
This section proposes to combine the wheelchair OAS operating system with a fuzzy control-based active user control strategy. Doing so improves the information collection rate, the information interaction effect, and the obstacle avoidance performance of the wheelchair. Specifically, the wheelchair motion control is realized by the user. The proposed automatic obstacle avoidance control strategy will guide the wheelchair to avoid obstacles successfully. Figure 7 displays the procedure flow of the proposed fuzzy automatic obstacle avoidance control strategy.
Based on Figure 7, the programming software codes the wheelchair OAA. Then, the OAA will be further tested through practical obstacle avoidance experiments.
3. Results
3.1. Information Interactive Performance Test of Wheelchair IIIS Operating System’s Directional Control
The experimental steps of the control and operation performance test of the wheelchair IIIS are as follows. First, the TSCS experiment is designed for the intelligent wheelchair IIIS. Then, the JCS experiment is carried out. The proposed wheelchair IIIS controls the front, rear, left, and right acceleration and deceleration of the wheelchair, respectively. SPHN TSCS interface is more suitable for elderly information acquisition. Specifically, it marks the four virtual buttons with prominent direction symbols. Then, it develops the acceleration, deceleration, emergency stop, braking, and whistle buttons. As such, it substantially improves the controllable interactive performance of the interface. Now that everything is ready, the TSCS and JCS control the wheelchair to accelerate, decelerate, and brake in the front, rear, left, and right directions, respectively. Figure 8 specifies the results.
Figure 8 indicates that the accuracy of SPHN TSCS is 98.1%, the average angle is 93°, and the total time cost is 54 seconds. The accuracy of the JCS is 99.2%, the average angle is 89°, and the total time cost is 50 seconds. Hence, the sensitivity of SPHN TSCS is higher than JCS; the joystick is more convenient to manipulate than the SPHN touch screen. The practicability of these two control methods is both very strong. Therefore, the proposed two independent control methods meet control performance requirements.
3.2. Angle Acquisition and Measurement
Mpu-6050 triple-axis gyroscope module is used to collect the inclination angle of the wheelchair. The theoretical measurement range is -180°~180°, and the accuracy is 0.01°. Here, the accuracy of angle acquisition and measurement is measured by absolute error (AERR). AERR is the difference between the analysis result and the actual value. The accuracy is usually expressed through error. The smaller the error is, the higher the accuracy of the analysis result is. Figure 9 displays the actual measurement results.
Figure 9 signifies that the maximum measurement error of the module is 0.3° and the minimum is -0.45°. Thus, the measurement accuracy is very high, and the functional performance is stable and reliable, meeting the design requirements. The obstacle avoidance control realized by the ultrasonic module can achieve a good obstacle avoidance effect in single and multiple obstacle environments. Ultrasonic obstacle avoidance has a larger obstacle avoidance angle and is safer than infrared obstacle avoidance. Therefore, the proposed fuzzy obstacle avoidance control strategy is feasible and meets the design requirements.
3.3. Simulation Analysis of OAA
The learning conditions in the obstacle avoidance simulation experiment are set as follows: the total number of moves per time is 1,000 steps, and the number of learning times is 20. Figure 10 illuminates the specific simulation effect and learning effect.
Figure 10 corroborates that during the learning process of automatic obstacle avoidance, the number of steps of autonomous walking of the intelligent wheelchair fluctuates significantly at the beginning and gradually decreases at the end. The steps of the obstacle avoidance movement also change greatly at the beginning and then slowly stabilize a little later. The learning effect is the ratio of the number of steps of autonomous walking to the number of steps of obstacle avoidance movement. The learning effect is significantly improved with the increase of learning times, so the obstacle avoidance performance of the proposed wheelchair IIIS has been improved.
4. Conclusion
Nowadays, there has been a rising voice for improving the quality of life of the elderly and the disabled. The present work holds that the elderly’s convenience to move is of utmost importance and practical significance. To this end, this paper designs a wheelchair IIIS based on fuzzy control theory and OAA. At the same time, a fuzzy obstacle avoidance control strategy is proposed for the wheelchair IIIS. Then, practical experiments are designed to test the performance of the proposed OAA of wheelchair IIIS. The experimental outcomes are as follows. (I) The accuracy is 98.1%, the average rotation angle is 93°, and the total time cost is 54 seconds for the TSCS. (II) For JCS, the accuracy is 99. 2%, the average angle is 89°, and the total time cost is 50 seconds. Hence, the SPHN TSCS is more sensitive than the JCS, and the control joystick is more convenient to manipulate than the touch screen in the SPHN for the elderly. The two independent control modes meet the control performance requirements, which improves the automatic obstacle avoidance performance of the wheelchair IIIS and has certain practical value. Based on their sensory functions, elderly users can obtain new experiences through different control methods. Also, they can enjoy barrier-free interaction between people and information equipment. The research deficiency is that the OAS experiment only involves static objects and lacks a dynamic performance test. Therefore, there is a need to continue to expand the use scene of the wheelchair IIIS in future research.
Data Availability
The labeled dataset used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The author declares no competing interests.
Acknowledgments
This research supported by the 2018 National Social Science Foundation Art Project No. 18BG128 and 2021 Humanities and Social Science Research Project of Henan Provincial Education Department: “Research on User-Oriented Interface Information Visualization for the Aging”, Project No. 2021-ZZJH-452.