Since the vibration energy is dissipated in the object on which communicating devices are placed and there are echoes from object boundaries as vibrations bounce back, a more sophisticated communication system design is required for VibeComm. The proposed method is ideal for low-rate off-line communication.
Independent and flexible technology
As its communication medium is the object on which devices are placed, communication over a longer distance and broadcasting to multiple devices are possible using VibeComm. These features distinguish the proposed method from [ 15 , 16 ] and radio-based communication methods. One possible application of VibeComm is the task of configuring a large number of devices simultaneously without using radios. For instance, consider a desk that is used by multiple users.
Suppose that a number of electronics, such as a notebook, a digital alarm clock, a tablet, an MP3 player, a telephone and a TV, are placed on the desk, and each electronic device on the desk has predefined user settings. If each user has her preferred settings for all electronics stored on her smartphone, then the proposed method can be used to configure all electronics on the desk for each user by simply placing the user's smartphone on the desk.
The remainder of this paper is organized as follows. An overview of the proposed method is shown in Section 2. Section 3 discusses issues when designing the proposed method, and Section 4 describes the design of VibeComm. The implementation of the proposed method is described in Section 5, and the results from the experiments are discussed in Section 6. Two-way and one-to-many communication demonstrations are given in Section 7, and remaining issues are discussed in Section 8. An overview of VibeComm is shown in Figure 1. We consider the problem of packet-based communication between smart devices that are equipped with accelerometers and vibrators.
The transmitting device takes an input from a user application and encodes the message into a series of packets. Each packet consists of frames, and each frame represents a bit of information. Each packet is transmitted from a transmitter by generating vibrations. A receiving device listens for incoming vibrations using its accelerometer and decodes incoming packets. The decoded message is then delivered to a user application. Due to the current hardware limitations discussed below, a sophisticated coding scheme cannot be applied.
A simple coding scheme is applied to VibeComm , where the presence of vibration in a frame indicates a bit 1 and the absence of vibration represents a bit 0. In order to reduce power consumption, impulse-like vibration signals are used to encode a packet, instead of continuous vibrations. In this paper, we focus on sending and receiving text messages to demonstrate the feasibility of vibration-based wireless communication.
In the next two sections, we discuss design issues, calibration of raw accelerometer readings and the design of the transmitter and receiver for VibeComm. An overview of VibeComm. A transmitter encodes a message into a series of vibrations. Vibrations propagate through the object on which the transmitter is placed. Receivers that are placed on the same object listen for incoming vibrations using accelerometers and decode them into messages.
Radio technology ensures reliable communications
We have encountered a number of issues while developing the proposed system, and they are time synchronization, inconsistent sampling frequencies, effects of echoes and detection sensitivity. VibeComm is designed to address these issues. Time synchronization between a sending device and a receiving device is critical for reliable communication. When time synchronization fails, a frame can be read incorrectly, and a bit error will result.
Unfortunately, the wall clock available in smart devices is not sufficient, since there can be an unknown length of delay from the propagation of the vibration signal. Hence, it is required to synchronize time between a transmitter and receivers each time communication starts for successful communication. In addition to time synchronization, we have observed that the sampling frequency of an accelerometer is inconsistent.
Furthermore, the sampling rate varies between models. We have also observed that both the sampling rate and the sampling interval vary depending on the load and state of the Android OS. Hence, it is not possible to sample with an exact sampling frequency with current Android platforms. An example is shown in Figure 2. We have transmitted a successive bit-stream of from the sending device and examined the received data from the receiving device.
Note that we have assumed that the sampling rate of the accelerometer is 17 Hz.
- Praying with Confidence: Aquinas on the Lords Prayer.
- Related Categories;
- Introduction to Supercritical Fluids: A Spreadsheet-based Approach!
- Absolute Zero.
The first seven bits are decoded correctly. However, the next seven bits are decoded as , which is a one-bit shifted version of the transmitted signal. This shows that a small misalignment can accumulate and cause a decoding error. A successive transmission of A bit error occurs due to time synchronization and an inconsistent sampling frequency.
As can be seen from Figure 2 , the received vibration signal is not located exactly in the middle of a frame, although it was transmitted in the middle of a frame. Actually, the location of the vibration signal drifts to the right as time passes. Since the direction and magnitude of the drift is not fixed, we have to regularly synchronize the clocks of the transmitter and receivers.
As the vibration signal propagates throughout the communication medium object, the signal can be reflected from boundaries and other objects placed on the communication medium. These reflected vibration signals can make decoding more challenging. We can reduce the effect of echoes by generating vibration for a short period of time. An example can be seen from Figure 2. In addition, this impulse-like vibration signal can reduce the power consumption of the transmitter.
While the effect of echoes was not considered in [ 15 , 16 ], it becomes a significant issue in VibeComm , as it is designed for communication over a longer distance through a rigid material. Since the vibration signal strength varies depending on the communication medium type the rigidness of the medium and decays as the communication distance increases, the signal detection sensitivity has to be set appropriately. Unexpected circumstantial factors, such as interfering vibration introduced by other objects, can make this problem even more difficult.
The problem resulting from the rigidness of the communication medium can be solved to some degree by adjusting the detection threshold. As shown in Figure 3 , the choice of the threshold value can make significant differences in decoded messages. The results shown in Figure 3 were obtained from a rigid wooden table, and the distance between the sending device and the receiving device was 50 cm.
See Section 4. As shown in Figure 3 , when the threshold gets smaller, the movement detection gets more sensitive. When the threshold gets larger, the movement detection is less sensitive and robust against disturbances. However, less sensitivity may cause the decoder to miss a signal from the transmitter.
Hence, the threshold adjustment based on the communication medium type and ambient disturbances is important for the success of VibeComm. In this section, the design of VibeComm is presented, including accelerometer calibration, the structure of a packet, the design of the transmitter and receiver, the design of a superframe, vibration signal shaping and adaptive threshold adjustment methods. In order to compensate for the hardware variation of an accelerometer in each smartphone, we need to convert raw accelerometer readings into the standard G unit using device-specific parameters.
We use the normalization scheme proposed in [ 17 ], which is required only once for each smartphone. A user is asked to hold his or her smartphone still and oriented towards different directions, which are not necessarily aligned with the direction of gravity.
After a sufficient number of samples are collected, the normalization parameters are estimated as follows. The following function is used in normalization:. When the phone is stationary, the function f is assumed to be one. Hence, to find the normalized accelerometer readings n x , n y and n z , we need to estimate parameters K x , K y , K z , b x , b y and b z , which make the function f unity when the phone is stationary.
In order to solve this parameter estimation problem, we use a least squares estimator based on the linear approximation of function f [ 17 ]. By doing this, we can minimize the error occurring due to the variation of devices. We consider packet-based communication in VibeComm , and a packet is contained in a superframe, a collection of frames. Each frame represents one bit of information. Figure 4 shows the structure of a superframe. In our current implementation, a superframe contains 11 frames. Each superframe starts with a beacon frame, which is used to synchronize devices. A superframe consists of a beacon frame one frame , active frames seven frames and inactive frames three frames.
The active frame field contains data, and seven frames are used in our current implementation to encode an ASCII character in each packet. The inactive frame field is a set of frames between the end of the active frame field and the start of the next superframe. While a lengthy inactive frame field is ideal in order to reduce the time synchronization error, we set the size of the inactive frames to be about a half of the size of the active frame field to improve the transmission rate. The effect of the inactive frame field will be discussed in detail in Section 4.
short-range wireless communication: Topics by diasabalreking.ga
In our current implementation, each frame is one second long. The resulting transmission rate is about 0. Currently, stable vibrations are not possible on Android platforms, and we have tested if the duration of a vibration is shorter than 0. Therefore, considering the resting time to minimize the effect of echoes, the minimum size of each frame suitable for reliable communication is experimentally found to be one second.
While the overall communication time is slow in our prototype, we believe that the problem will be addressed when more sophisticated and reliable vibrators are available in future smartphones. In order to transmit a bit 1, we generate an impulse-like vibration for 0. No vibration in a frame means a transmission of 0. The inactive frames do not contain vibrations, and a beacon frame also transmits a 0.
This impulse-like vibration signal is used to reduce the power consumption of the transmitter and to avoid the effect of echoes. The value f in Equation 1 represents the magnitude of a normalized accelerometer reading. In order to determine the existence of an incoming vibration signal, the magnitude of a normalized accelerometer reading is compared to the sample standard deviation.