key: cord-0034674-ixjfcy93 authors: Zeiß, Steeven; Marinc, Alexander; Braun, Andreas; Große-Puppendahl, Tobias; Beck, Sebastian title: A Gesture-Based Door Control Using Capacitive Sensors date: 2014 journal: Distributed, Ambient, and Pervasive Interactions DOI: 10.1007/978-3-319-07788-8_20 sha: 2836627907c0246287b1fffe3deb8271db5850e6 doc_id: 34674 cord_uid: ixjfcy93 In public places sanitary conditions are always of concern, particularly of surfaces that are touched by a multitude of persons, such as door handles in rest rooms. Similar issues also arise in medical facilities. Doors that open based on presence are common in environments such as shopping malls; however they are not suited for sensitive areas, such as toilet stalls. Capacitive proximity sensors detect the presence of the human body over a distance and can be unobtrusively applied in order to enable hidden gesture-based interfaces that work without touch. In this paper we present a concept for a gesture controlled automated door based on this sensor technology. We introduce the underlying technology and present the concept and electronic components used in detail. Novel interaction patterns and data processing methods allow to open, close, lock and unlock the door using simple gestures. A prototype device has been created and evaluated in a user study. The importance of proper sanitation is a well-understood principle to help prevent the spreading of diseases. Door handles in public spaces can be used by hundreds of people in the scope of the day and may act as a source of infection by passing bacteria or different virus between persons [1] . Ideally, these surfaces should either be cleaned or contact be avoided. Automated doors are commonplace in modern environments, e.g. in front of shops, to allow easy entrance and minimize heating cost. However, these are limited to detecting the presence of persons and usually rely on a simple timer to close again. This prevents application in scenarios where the door requires more than one mode of operation, e.g. if it has to be locked and unlocked. In this paper we present a method for a gesture-based door control relying on a set of capacitive sensors that are able to detect the human body over a distance. Using a few simple hand gestures in front of an unobtrusively applicable box it is possible to control an automated door without touching any surface, thus considerable reducing associated health risks. We consider three different potential applications. Public toilets in crowded areas can be visited by a large number of persons each day and have numerous lockable stalls. There are various technologies that improve sanitation, including automated flushing and self-cleaning seats, controlled using presence sensors. However, in order to lock and unlock the stalls it is still necessary to use a handle, as mere presence sensors do not have the required expressiveness to control the locking process. This can be achieved using a set of gestures and our proposed system. The second application area is hospitals. Sanitation is a major concern here and often doors have to be controlled without the use of hands, e.g. using foot switches. The automated doors typically only allow opening and self-close after a preset amount of time. Using our system it is possible to use dedicated opening and close gestures that could be also controlled with a foot-based system. A final application would be door controls for persons with physical disabilities. When there are limitations to the fine motoric skills of the user it is possible to configure the system in a way that it can be controlled by coarse gestures that are specifically tailored to a user. On the following pages we will briefly introduce the technological basis of capacitive proximity sensors and discuss the related works before detailing the system design and gesture sets available to control the system. We will describe the components of the prototype we created, including components used. Finally, the results of a user evaluation testing the usability of the system will be presented. Capacitive proximity sensing is a fairly old technology, first introduced by Russian physicist Leon Theremin around 1918, who created an early electronic instrumentthe eponymous Theremin [2] . It allows controlling pitch and volume of a generated sine wave by moving the hands in the range of two distinct antennas. Thracker prototype attached to a display [3] The potential applications in HCI have been a research interest at the MIT in the 1990s [4] . Additionally, the technology has been used for different interaction devices in the last years [5, 6] . A closely related application is Thracker, developed by Wimmer et al. and shown in Figure 1 [3] . Using four sensors that are placed on the corners of a monitor it is possible to track the position of a hand moving in front of it and detect grasp gestures, in order to control typical UI functions, such as scrolling and zooming. Capacitive sensors measure the capacitance of an electric system. Both the sensor and the human body act as a part of an electric field that is generated with regard to a common ground. Using a simple plate condenser model this relationship can be described using the following equations for capacitance C, charge Q, electrode area A, distance between plates d, vacuum permittivity , dielectric constant and electric field strength E [7] . , Smith et al. distinguish three different measurement modes of capacitive proximity sensing of the human body [8, 9] . Transmit mode uses a dedicated sender and receiver electrode, whereas the sender is placed close to body creating a capacitive coupling, resulting the body to act as sending electrode. This allows the sender and receiver to be placed further apart. The second mode is shunt mode using a field created between a dedicated sender electrode and a dedicated receiver electrode. The human body entering this field causes a displacement current, reducing the overall capacitance that can be associated to a distance from the center point between the two electrodes. The final method is loading mode that uses a single electrode creating an oscillating electric field with regards to the environmental ground. If the human body enters this area, the capacitance of the system increases with regards to the proximity between electrode and body. The latter is used as method of choice for our proposed system. Loading mode allows for a simpler technical setup and increased detection distances on a planar and dense electrode layout. As described in the previous sections, capacitive proximity sensors are employed when hand gestures close to the door have to be recognized. For a simple-to-build and low-cost setup, four sensors are sufficient and provide basic gesture recognition. These sensors allow for recognition of circular gestures and horizontal or vertical movements, also known as swipe gestures. The four sensor electrodes are arranged in a diamond shape, as depicted in Figure 2 . For swipe gesture detection, only two sensors will be active: either the horizontal, or the vertical ones. The door lock is installed as a stand-alone, as it can be opened manually from one side only. The general process of recognizing gestures is depicted in a state diagram, shown in Figure 3 . The microcontroller stays idle until one of the sensors' measurements exceeds a threshold value. Once that has occurred, the sensor's current ID is written into a buffer. Depending on the start-and endpoints of a gesture, e.g. when leaving and entering the interaction area, the buffer's values are matched to pre-defined sequences, similar to Dynamic Time Warping. In case no pattern was recognized, the microcontroller is turned back into idle mode and waits for the next gesture trigger. When a pattern has been recognized successfully, succeeding actions can be executed -for example opening of the door. In order to recognize gestures, sets of predefined patterns are defined a priori. These patterns contain the sequential activation orders of the four sensors. As a simple example, pattern 12341 would represent a circular gesture, whereas pattern 1111444 would represent a vertical swipe gesture. A door can be reduced to locking. We diagramed the from left to right, and vice v from left to right, and vice matching these gestures to t ch, circular gestures are mapped to locking and unlock as developed in analogy to ordinary doors, as keys or d nd to lock and unlock the door. Moreover, the direction lso given by a natural mapping and habits. Thus, we ar ve approach, and will prove this assumption in our eval pening and closing the door is initiated by a horizonta plained in method 1, this bears the danger of opening nt when there is limited space around the gesture recog oor is locked, an open gesture can be restricted: The d r to opening it. Figure 4 depicts the two different meth user study. Locking and Unlocking t does not depend on a dedi opened by manually turnin tures. To keep the option people find it hard to cope tricity blackouts and evacua realized in software with th attached to the gesture reco really necessary, we argue intention of people feeling to extend the setup by an a This would allow for additi Opening and Closing the tor-control with two input li pulled to 3.3V. In order to follow during the next seco the actual door, as it is not or closed). Therefore, an in trigger a different system be We have created a prototyp tem with the door applied o into a single unit that only Implementation the Door. As described in the previous section, the se cated door lock. When the door is closed, it may only ng the door knob on the inside of the restroom, or by g of manually opening the door is very important in c with gesture recognition systems, or in the event of el ations within a building. Locking and unlocking the doo he possibility of indicating the current state by a status L ognition device. Even though locking and unlocking is that the psychological effect cannot be neglected with more secure when the door is locked. It is also imagina additional capacitive sensor on the outer part of the do onal security when the door is opened after unlocking it Door. In the previous sections, we described the door m ines for mechanical buttons. Therefore, the enable signa o trigger the motor control, two succeeding impulses m ond. Unfortunately, there is no way to distinguish betw t communicated to the peripheral components (either o nternal marker is used that saves the current state and m ehavior. pe of the system, as shown in Figure 5 . It is a portable s on a stand. Sensor and evaluation electronics are integra y requires a power supply and a single connection to Fig. 5 . Overview of the prototy ing mechanism. Right -contro The basic sensor is based on capacitance of the system. T on the IC TLC555 provided the OpenCapSense rapid p oped by Grosse-Puppendah n an oscillating circuit that changes frequency based on The layout of the board is shown in on the left and is ba d by Texas Instruments [10] . The system is grounded up prototyping toolkit for capacitive proximity sensing dev hl et al. [11] . ut of sensor module. Right: layer model of sensor electrode olkit include different filtering algorithms, such as mov g with a variable sample count. In this project we us mean and moving average to generate a smooth signal onally, we employ a guard electrode to prevent exter ng the measurements. In order to control the diffe unit and an input channel o of the eight provided sensor free channels to control a r We employ a FTR-B4 prov tion and can be controlled u We have evaluated the usa was placed at a height of ab to compare both of the pre tem concept. There were 1 ages 21 and 35 (mean 27). avoid learning bias. Afterw various Likert-scale questi There is a visible and fast le the users were trying out th generally lower, implying a for any set, however, circul Related to questionnaire 7.13, which was somewhat ble swiftly. This can be att Many users tried to access t some questions regarding th to be very important in thos a system such as this to be that the current hygienic sta The questionnaire also i improvements of the system visual feedback of the curr was not considered sufficie determined whether an an erent functions of the door, an interface between the con of the motor control is used. As we are only in need of f r channels of OpenCapSense it is possible to use one of relay attached to the door switches, as shown in Figure vided by Fujitsu-Takamisawa that has low power consum using the board supply of 5 V. Circuit diagram of the simulated switch ability of the system using our prototype. The control u bout 1.20 m on the left part of the door frame. We wan sented gesture sets and get a general feedback on the s 16 users participating in the study (2 female) between They were randomly assigned an order of the two set wards the users had to fill out a small questionnaire w ions (scale 1-10). The error results are shown in Fig earning effect from first to last gesture regardless of set he system at this point. The error rate on the second set w a significant learning effect. There is no obvious prefere ar gestures were considered more intuitive. results, the perceived speed was ranked with an average below expectations as the system was designed to be u tributed to the higher error rate when performing gestu the system from further away than possible. We also as he use case of public toilets. The users considered hygi se areas, with a rating of 8 In this paper we presented a gesture-based door control system based on capacitive proximity sensors that allows controlling doors in public spaces without touching any surface. We have provided the system concept and two sets of gestures that can be used to control the door. The system was implemented in prototypical form and evaluated in a study with 16 participants. The results show that the system is intuitive to use and that the subjects would strongly prefer this system to touching door handles in public toilets, indicating that this system is a viable alternative to current solutions. However, our system signifies only a first step in this direction and there are numerous improvements that can be applied to future iterations. The door system we use does not provide its current state to our control unit, thus it has to be tracked with software. In the future, we plan on switching to a system providing this functionality, or integrate additional sensors that measure the state. Another addition to the system, suggested by study participants, is providing a better visual feedback. It is particularly interesting to display whether a door is closed or open. We are considering using a LCD display that could provide an iconic representation of the system state if an error has occurred or if the hand is obstructing the gesture area. This feedback should ideally be available on both sides of the door. Finally, we would like to investigate other interaction systems that provide an even higher expressiveness. The Leap Motion allows for a fine detection of finger and hand locations [12] . This enables detecting actual gestures associated with doors, such as grabbing a virtual door handle or moving an imaginary key. Particularly for user groups that are often using these gestures, this might increase the intuitiveness and acceptance. Effects of cleaning and disinfection in reducing the spread of Norovirus contamination via environmental surfaces Theremin: Ether music and espionage Thracker -Using Capacitive Sensing for Gesture Recognition Applying electric field sensing to human-computer interfaces Using the human body field as a medium for natural interaction Honeyfish -A high resolution gesture recognition system based on capacitive proximity sensing Electric field imaging Electric field sensing for graphical interfaces Elektronik-Fibel. Elektron. Bauelemente, Schaltungstechnik, Digit OpenCap-Sense: A Rapid Prototyping Toolkit for Pervasive Interaction Using Capacitive Sensing