Existing non-image-based solutions for detecting Phalaenopsis orchid diseases rely on time domain analysis of environmental conditions. These solutions are suitable for loosely controlled open farm fields. Conversely, the diseases are infrequent events in Internet of Things (IoT)-based smart orchid greenhouses, where environmental conditions are well-controlled. In tightly controlled settings, standard time domain methods fall short in detecting subtle changes in environmental conditions during rare orchid disease events. We observed that the influence of the time series of temperature and humidity on orchid fungus within the well-controlled greenhouse is not readily apparent, and the time-frequency coefficients of temperature and humidity must be used to detect orchid disease in this rare event case. In order to address this challenge, we propose OrchidTalk, an IoT-based deep learning platform, which employs a time-frequency domain analysis using the Continuous Wavelet Transform (CWT) filter, for the purpose of detecting rare events of orchid diseases. Through a non-trivial process, we select the Gaussian wavelet with four derivatives to serve as the mother wavelet of the CWT in the data extraction phase, and invent the CWT IoT device to translate the time-domain IoT raw data to the time-frequency domain IoT data. The resulting CWT features are used as inputs to the Orchid-3D model (three-dimension ConvLSTM), which significantly outperforms the previously proposed Orchid-1D (CNN and LSTM) and Orchid-2D (two-dimension ConvLSTM) models. For the rare event detection of 0.2% of sick orchid plants, OrchidTalk achieves a recall of 0.902, precision of 0.92, accuracy of 0.99, and an F1-score of 0.911. To the best of our knowledge, this represents the highest prediction performance achieved for a well-controlled orchid greenhouse.

The Bacillus genus is one of the most commercially exploited bacteria in the agro-biotechnology industry, and the Bacillus information is very useful for crop growth. Most existing studies on the analysis of the amount of Bacillus were conducted in laboratories. Performing such a task on open field farming is difficult because only a small data set is available during a long observation period for the soil analysis of Bacillus. For example, turmeric growth takes nine months with one soil sample per month, and we found that increasing the frequency of soil analysis for turmeric growth is not practically useful. Therefore, we can only collect a very small data set for AI training. This article proposes the AgriTalk approach that predicts the amount of Bacillus based on novel IoT and machine learning technologies. AgriTalk uses a small data set (five data items) per farm for training and performs prediction for the subsequent four months. Good results are obtained. Specifically, the inference mean absolute percentage errors (MAPEs) range from 6.73% to 19.76%. In the experiments of five farm fields, we have correctly captured the trends for the number of changes of Bacillus. Such prediction provides useful information for fertilization management. Our prediction is more accurate for farms covered by peanut shells (the average MAPE is 13.24%) than for farms covered by rice husks (the average MAPE is 15.43%).

Air and soil sensors are used in smart agriculture to autonomously collect the soil-related data. It is important that the measurements of the sensors are accurate. This article uses electrical conductivity (EC) as an example to show how to avoid the inaccuracy due to EC sensor aging. We consider a promising solution that packagesIredundant EC sensors in one device (${I}>{1}$) called${I}$-redundant EC sensor package. The redundant-sensor package outputs the measured data based on a voting process among${I}$sensors, which makes soil measurement more accurate and reliable. Based on the${I}$-redundant EC sensor package, this article invents the first out-of-the-laboratory automatic calibration procedure in the farm field. We also propose the SensorTalk mechanism to extend the life of the${I}$-redundant EC sensor package through calibration. Most people have the illusion that the life of a redundant-sensor package is longer than a nonredundant sensor. This article proves that the life of a redundant-sensor package is shorter than its nonredundant counterpart. Then, we propose the self-calibration process of the redundant-sensor package to extend its operation life. For example, if the life of a sensor in the${I}$-redundant package is six months with the aging-after-calibration factor${\gamma }={0}.{9}$, then SensorTalk may extend its life to 52.2 months for${I}={9}$. We also shed some nontrivial insights. In many smart farming solutions, the sensors typically report the measured data with a frequency lower than one per 10 min. Our study indicates that by increasing the sampling frequency up to one per 20 s, calibration for an EC value can be sped up by 19 times. Our study also suggests that an oddInumber should be chosen. If we add one more sensor to make an even${I}$number, the life of the redundant-sensor package is actually shortened.

Sensor failure detection and calibration is an important issue for Internet of Things applications. Many smart applications fail due to inaccurate data produced by their sensors. To solve this issue, we propose the CalibrationTalk mechanism. We use smart agriculture as an example to show how CalibrationTalk automatically fixes the sensing accuracy problems in commercial farm operations. Besides manufacturing variation, hardware malfunction and aging, sensing inaccuracy may be caused by soil erosion of rain or irrigation, which changes the positions and the angles of the sensors inserted in the soil. The above sensor failure problems can be detected by CalibrationTalk. In this article, we first extend our previous work to detect sensor failures. If the sensors become inaccurate due to aging or similar reasons, CalibrationTalk automatically calibrates the sensors in the farm fields. We conduct measurements and analytic modeling to investigate the performance of CalibrationTalk, which suggests that to detect a potential moisture sensor failure, it is appropriate to set the detection time as 30 min for moisture sensors. With this setup, true failure is detected while false detections rarely occur. For a detected sensor failure due to aging, CalibrationTalk can automatically build a calibration table. During the calibration table establishment period, the sensors still produce correct measured data and do not interfere with the normal operation of smart farming. The value of CalibrationTalk lies in its power of automation for sensor failure detection and calibration processes developed in this article.

Internet of Things (IoT) integrates with finger tracking and 3D signature verification can provide the user authentication service for these IoT applications which require user identification and authorization, such as the home automation and security system. This work develops two modules. The FingerTalk module provides the fingertip trajectories tracking service based on a Leap Motion camera and IoTtalk, The AirSig module provides 3D signature verification service to recognize whether a signature is signed form the authenticated user. The AirSig module requires only two user signature samples to train the neural network model which significantly reduces the process time for signature collection. The validation results from 10 people show the average accuracy is up to 93.6% on 3D signature verification.

Rice blast is one of the most serious plant diseases. Many rice blast management approaches require know-how of experienced farmers or agronomists. Monitoring the farm for disease detection is labor intensive and time consuming. By using the Internet of Things (IoT) and artificial intelligence (AI), we are able to detect plant diseases more efficiently. Existing AI and IoT studies detect plant diseases by images or nonimage hyperspectral data, which require manual operations to obtain the photographs or data for analysis. Also, image detection typically is too late as rice blast may already spread to other plants. Based on an IoT platform for soil cultivation, we develop the RiceTalk project that utilizes nonimage IoT devices to detect rice blast. Unlike the image-based plant disease detection approaches, our agriculture sensors generate nonimage data that can be automatically trained and analyzed by the AI mechanism in real time. The beauty of RiceTalk is that the AI model is treated as an IoT device and is managed like other IoT devices. In this way, our approach significantly reduces the platform management cost to provide real-time training and predictions. We also propose an innovative spore germination mechanism as a new feature extraction model for agriculture. In the current implementation, the accuracy of the RiceTalk prediction on rice blast is 89.4%.

Internet of Things (IoT) has been widely utilized with artificial intelligence (AI). However, it requires substantial effort to integrate AI and big data with IoT. To mitigate this problem, AItalk was proposed. By treating AI as a cyber IoT device, we do not need to write code of the AI mechanism in the network applications, as traditional AI-based IoT applications did. This article describes how the AI tools such as scikit-learn and TensorLayer are accommodated as cyber IoT devices in AItalk, and extends AItalk for non-IoT applications. We use house valuation as an example to show how AItalk can flexibly include the factors that have significant impact on the house price. We show that by adding extra features other than housing profile features, the accuracy for the prediction (valuation) can be improved by 38%. We also investigated the communication overhead of the distributed AItalk structure, which is 3.7% for the computation of one house price valuation.

This paper proposes AgriTalk, an inexpensive Internet of Things (IoT) platform for precision farming of soil cultivation. We conduct experiments on turmeric cultivation, which indicates that the turmeric quality is significantly enhanced through AgriTalk. Specifically, the curcumin concentration is up to 4500-5500 mg/100 g, which is five times more than existing products. We demonstrate how to intuitively configure the connections between the sensors and the actuators with the desired farming intelligence, and to effectively maintain AgriTalk for precision farming. We conduct measurement, analytic analysis, and simulation experiments to investigate the IoT message delays of AgriTalk. This paper indicates that the delays for automatic control and automatic-manual control switching with long distances (more than 30 km) are very short (less than 0.2 s) and AgriTalk can easily respond to quick and dynamic change of the field environment conditions in soil cultivation.

Narrowband Internet of Things (NB-IoT) is considered as a promising wireless communications technology for Internet of Things (IoT) especially for the outdoor environment. Many outdoor IoT applications involve large numbers of homogeneous NB-IoT devices. It is tedious to specify and accommodate these devices during application development. To resolve this issue, this paper proposes a service platform for fast development of NB-IoT applications called NB-IoTtalk. This platform utilizes a tag mechanism to provide an easy-to-manipulate graphical user interface to accommodate a large number of NB-IoT devices in an application and transparently show them in a visual map. Our approach automatically creates and parses the device profile used to interpret the payload of an NB-IoT message. We then use a smart parking lot application as an example to investigate the event-triggered reporting of NB-IoT in terms of the time-to-live (TTL) report frequency and the outage detection accuracy. This paper provides the guidelines to set the TTL interval for event-triggered NB-IoT applications.

Multi-sensorial effect or four-dimensional (4-D) effect is considered in the niche arena of film playing but has become a growing business quickly. Installation of 4-D effects are typically expensive, and 4-D films are most often presented in custom-built theaters at special venues. This paper proposes a novel integration of existing video service platform and commercial smart home solution to generate 4-D effects for any non-4-D film (i.e., a 3-D or a 2-D film). Specifically, we describe how the designer can use home appliances to create the effects for a 4-D experience theater at home. Instead of using an expensive and special 4-D movie effect system, we use a low-cost off-the-shelf IoT platform for a smart home to create the 4-D movie effects for any non-4-D film. Therefore, the viewer can enjoy 4-D movies at home without special equipment other than home appliances. In our solution, an IoT device-management system called IoTtalk is utilized to integrate the video service system and the smart home system. Since these systems may be located in distant places, the delays for video streaming and triggering of 4-D effects must be synchronized. We have conducted measurements to indicate that in our design, the synchronization issue can be ignored without affecting the user experience.

IoTtalk is a platform for Internet of Things (IoT) device interaction, which nicely integrates a reconfigurable multi-sensor device called MorSensor with the proposed IoT management platform in the network domain. The sensors can be dynamically plugged in/out of a MorSensor device without being turned off, and IoTtalk automatically generates/reuses the application software for these sensors. We propose a dynamic ranging concept that automatically specifies the value range of a sensor so that it can send “meaningful” data to any connected output IoT device. Therefore, a MorSensor device can “talk” to other IoT devices, such as a light bulb or an electric fan. We also show that IoTtalk is a simple yet almost free solution for automatic sensor calibration. Finally, we illustrate that IoTtalk is a powerful tool for developing interactive science experiments.

Almost half of all Americans either own smart home technology or plan to invest in it. Therefore, many commercial smart IoT solutions aim for home automation, and try to live up with high expectation for smart home. A major problem of these smart home solutions is that they cannot flexibly accommodate the existing remotely controlled aftermarket appliances. This paper proposes IoTtalk remote control (IoTtalk-RC), a mechanism that utilizes sensors as universal software-defined remote control for aftermarket home appliances. Our solution replaces the infrared remote controller of cheap aftermarket home appliances and provides extra intelligence of sensors that can be manipulated by the homeowners. We conduct measurements, analytic analysis and simulation experiments to investigate the time complexity of IoTtalk-RC. This paper indicates that the extra transmission delays of remote control incurred by IoTtalk-RC are small even if the IoTtalk server is remotely located at a virtual machine of a cloud. In other words, IoTtalk-RC is feasible and can be effectively operated with the IoT technology.

In this paper, we study on the price comparison between shared plan and non-shared plan for mobile service. A non-shared data plan is a traditional data plan for a single user with a limited monthly data allowance. A shared data plan allows a user to add different devices to one inclusive plan and receive one bill at the end of every month, which the monthly data allowance is shared by a group of users across multiple devices. Each data plan is charged by a fixed price for a monthly data allowance and there will be an extra charge if a user consumes more than the monthly data allowance. Therefore, the charged price is affected by the user's behaviors. We study the users' behaviors by observing the measured data from Chunghwa Telecom and derive stochastic models for shared plan and non-shared plan to evaluate monthly prices based on different users' behaviors. We recommend the most appropriate plan for a group of users based on users' behaviors. On the other hand, mobile operators can attract more customers and earn more profit. We show that the shared plan is a `win-win' plan for users and operators.

Many Internet of Things (IoT) technologies have been used in applications for money flow, logistics flow, people flow, interactive art design, and so on. To manage these increasing disparate devices and connectivity options, ETSI has specified end-to-end machine-to-machine (M2M) system architecture for IoT applications. Based on this architecture, we develop an IoT EasyConnect system to manage IoT devices. In our approach, an IoT device is characterized by its “features” (e.g., temperature, vibration, and display) that are manipulated by the network applications. If a network application handles the individual device features independently, then we can write a software module for each device feature, and the network application can be simply constructed by including these brick-like device feature modules. Based on the concept of device feature, brick-like software modules can provide simple and efficient mechanism to develop IoT device applications and interactions.

A ticket dispenser (TD) is used to assist customers for the waiting process in, e.g., a shop. This paper deploys a mobile ticket dispenser system (MTDS) with waiting time prediction to enhance user experience in waiting. For example, the MTDS for a restaurant allows a customer to remotely draw a ticket for meal order anywhere through a smart phone before she/he arrives at the restaurant and therefore reduces her/his waiting time. We propose an output indicator and develop a discrete event simulation model to investigate the performance of the MTDS. Our study indicates that the waiting times can be more accurately predicted without consuming much wireless network resources and power consumption of mobile devices.

A mobicast, or called as mobile geocast, problem in three-dimensional (3D) underwater sensor networks (USNs) is investigated in this paper, which aims to overcome the hole problem and minimizes the energy consumption of the sensor nodes while maximizing the data collection. In this work, all underwater sensor nodes are randomly distributed in a 3D underwater environment in the sea to form a 3D USN. Considered a mobile sink or an AUV (autonomous underwater vehicle), all possible sensor nodes near to the AUV form a 3D geographic zone (called as 3D zone of reference or 3D ZOR). The AUV travels a user-defined route and continuously collects data form sensor nodes within a series of 3D ZORs at different time. The main problem is how to efficiently collect data from sensor nodes within a 3D ZOR while those sensor nodes are usually in sleep mode for a long period of time. The routing protocol relies on two phases; the first phase is to collect data form sensor nodes within a 3D ZOR, and the second phase is to wake up those sensor nodes in the next 3D ZOR to be queried while trying to avoid topology holes. To save power, only sensor nodes in a 3D ZOR are notified to enter the active mode in order to deliver sensed results to the AUV. To consider the characteristics of USNs, a new mobicast routing protocol is developed in 3D USNs. The key design challenge is to develop a power-saving mobicast protocol in 3D USNs to overcome the unpredictable 3D hole problem. An ”apple slice” technique is used to build multiple segments to surround a hole and to assure routing path continuity. Finally, performance analysis is derived and simulation results illustrate the performance improvement in successful delivery rate and power consumption.

In this paper, we present a spatiotemporal multicast, called a mobicast, protocol for supporting applications which require spatiotemporal coordination in VANETs. The spatiotemporal character of a mobicast is to forward a mobicast message to vehicles located in some geographic zone at time t, where the geographic zone is denoted as zone of relevance (ZOR). Vehicles located in ZOR at the time t must keep the connectivity to maintain the real-time data communication between all vehicles in ZOR. The temporal network fragmentation problem is occurred if the connectivity of ZOR is lost such that vehicles in ZOR cannot successfully receive the mobicast messages. To solve the problem, a new mobicast protocol is presented in this work to successfully disseminate mobicast messages to all vehicles in ZOR via a special geographic zone, called as zone of forwarding (ZOF). The main contribution of this work is to develop a new mobicast routing protocol to dynamically estimate the accurate ZOF to successfully disseminate mobicast messages to all vehicles in ZOR. To illustrate the performance achievement, simulation results are examined in terms of dissemination successful rate, packet overhead multiplication, and packet delivery delay.

This paper investigates a new TDMA-based MAC, called C-MAC (MAC using Chinese remainder theorem), protocol with low power consumption while maintaining low packet latency for large-scale cluster-based WSNs. To offer low power consumption, each node and coordinator to be active and sleep states based on the time-slot schedule of Chinese remainder theorem. To provide low packet latency, our C-MAC protocol provides an adaptive time-slot scheme to distributively and dynamically wake up time slots for the burst data transmission in a duty cycle. Finally, simulation results illustrate our performance achievements to verify that C-MAC performs better than existing TDMA-based MAC protocols, BMA and EMAC, in terms of power consumption and average packet latency.

The quality-of-service (QoS) communication that supports mobile applications to guarantee bandwidth utilization is an important issue for Bluetooth wireless personal area networks (WPANs). In this paper, we address the problem of on-demand QoS routing with interpiconet scheduling in Bluetooth WPANs. A credit-based QoS (CQ) routing protocol is developed which considers different Bluetooth packet types because of different types of Bluetooth packets have different bandwidth utilization levels. This work improves the bandwidth utilization of Bluetooth scatternets by providing a new interpiconet scheduling scheme. Centralized and distributed algorithms are investigated in this work to improve the bandwidth utilization for the on-demand QoS routing protocol The performance analysis illustrates that our credit-based QoS routing protocol achieves enhanced performances, compared to existing QoS routing protocols.