In this article, we design a deep learning based pedestrian dead reckoning (DL-PDR) framework with vision anchor augmentation for seamless centimeter-level indoor positioning. The DL-PDR framework integrates deep learning with PDR to address the inherent inaccuracies associated with estimating step lengths and heading directions in PDR methods, as errors in these estimations propagate and accumulate over time, resulting in progressively larger positioning errors. In addition, DL-PDR explores vision-based positioning using existing surveillance cameras while incorporating dynamic anchor augmentation to achieve high-precision real-time localization and correction. According to our review of relevant research, this is the first framework that explores and integrates vision anchor augmentation with inertial data recognition in a deep learning manner to accurately recognize individual walking distances, heading angles, and up-to-date locations, seamless achieving positioning accuracy at the centimeter level. We have implemented an Android-based system with network cameras to validate the feasibility and superiority of the DL-PDR framework. Experimental results show that our framework outperforms existing methods and can accurately recognize the walking distances, moving angles, and located points of individuals with centimeter-level accuracy.

This paper designs and implements a head-mounted individual monitoring and finding system, called iHead, to achieve proactive individual safety and assistance services through Internet of Things (IoT) technologies. The iHead system utilizes inertial data to monitor individual behavior and detects incidents (e.g., falls, crashing accidents, head hitting, etc.) based on deep learning. The individual activity recognition model is developed based on heterogeneous learning attention of long short-term memory, feedforward neural networks, and densely connected neural networks to accurately identify individual status. In addition, in daily use, iHead can recognize bad sitting posture and provide suggestions for correcting the sitting posture. Furthermore, in hazardous situations, iHead can automatically issue alerts and actively track the target individual in a crowdsourced sensing manner. To validate the feasibility of iHead, an Android-based prototype with the head wearable device for individual monitoring and finding has been implemented.

This article proposes a proactive crowdsourced monitoring and sensing (PCMS) framework with the designed Smart iBeacon device to accurately recognize the activities of an equipped target, exclusively customize the recognition model of a specific target, and actively trigger cooperative tracking of nearby smartphones for an abnormal target based on Internet of Things (IoT) localization. According to our review of relevant research, PCMS is the first framework that provides the following features: 1) coarse-grained and fine-grained features can be extracted to accurately recognize target activities through densely connected convolutional networks with improvement design; 2) crowdsourced monitoring and sensing can be actively triggered for a target as the abnormal activity of the target is detected; and 3) deep learning model of target activity recognition can be exclusively customized for a specific target to improve the recognition accuracy based on the dedicated activity data of the target. An Android-based prototype with stationary iBeacon nodes and the Smart iBeacon is implemented to verify the feasibility and superiority of our PCMS framework. Experimental results show that our framework outperforms existing methods and can accurately recognize target activities for abnormal event detection and proactive crowdsourced tracking in a real-time manner.

The popular East Asian card gameBig2involves rules that do not allow players to view each other's hand cards, making artificial intelligence face challenges in performing well in this game. Based on Markov decision processes (MDPs) that can handle partially observable and stochastic information, we design the Big2MDP framework to explore card-playing strategies that minimize losing risks while maximizing scoring opportunities for theBig2game. According to our review of relevant research, this is the firstBig2artificial intelligence framework with the following features: first, the ability to simultaneously consider scoring and losing points to make the best winning decisions with minimal losing risk, second, the capability to predict multiple opponents' actions to optimize the decision-making, and third, the adaptability to compete for the free-playing right to change card combinations at the essential moment. We implement a system of four-player card gameBig2on the Android platform to validate the feasibility and effectiveness of Big2MDP. Experimental results show that Big2MDP outperforms existing artificial intelligence methods, achieving the highest win rate and the least number of losing points as competing against both computer and human players inBig2games.

In this paper, we propose a cooperative fleet lane changing (CFLC) framework based on vehicular sensor networks, which enables multiple groups of fleet vehicles splitting and merging. The proposed CFLC framework determines the lane-changing space in the target lane, performs real-time splitting of fleet vehicles, coordinates the speeds of fleet vehicles and non-fleet vehicles, and merges multiple groups of fleet vehicles. In CFLC, we design the dynamic programming algorithm to determine the optimal fleet lane-changing space, and split fleet vehicles into multiple groups according to the optimal lane-changing space. As there are no sufficient inter-vehicle gaps for fleet group lane changing, hybrid speed control is explored to adjust both the fleet vehicle speeds and the speeds of vehicles in the target lane to avoid unnecessary acceleration/deceleration. Moreover, efficient merging of multiple fleet groups is designed to deal with the separation of fleet vehicles for avoiding increased travel time and fuel/energy consumption. According to our review of relevant research, this is the first cooperative fleet lane changing framework that supports multi-group vehicle splitting and merging. Simulation results show that our framework outperforms existing lane-changing methods and can minimize the speed variations of fleet vehicles and non-fleet vehicles for improving traffic flow and reducing fuel/energy wastage. In particular, with a fleet of 10 vehicles at 110 km/hr, average speed variations for changing lanes are decreased by more than 14 m/s and improved by more than 45%.

This paper designs and implements a rapid shopping navigation system, called NaviEyes, to minimize shopping time and maximize purchasing success ratio for hypermarket customers through Internet of Things (IoT) localization. The NaviEyes system can provide a shopping trip with the shortest total moving time to purchase all desired products according the crowd density of each passageway/area. In addition, NaviEyes can plan the proper shopping order and fast moving paths between selected on-sale products to minimize the purchasing risk (i.e., one or more selected on-sale products sold out) through avoiding taking long moving time to purchase a selected product with sufficient instances. An Android-based prototype with video-based indoor localization is implemented to verify the feasibility and performance of our NaviEyes system.

This study proposes a centimeter-level indoor positioning (CLIP) framework to achieve highly accurate localization with facing direction detection for the microlocation-aware Internet of Things (IoT). The CLIP framework can provide accurate centimeter-level positioning information to people indoors by integrating installed surveillance cameras with the IoT, where the efficient operation of microlocation-aware IoT applications and services can be enabled for smart spaces. CLIP can be used to accurately determine the position and facing direction of an individual. According to our review of relevant research, CLIP is the first indoor positioning framework that includes the following features: 1) centimeter-level positioning accuracy for the microlocation-aware IoT that can detect the facing direction of individuals; 2) employment of existing surveillance cameras with low-additional installation cost; and 3) innovative infrastructure for microlocation-aware IoT applications that can enable accurate centimeter-level path planning for individuals, emergency evacuation for groups of people, and geofencing with microlocation awareness. An Android-based system was implemented to verify the feasibility and effectiveness of the CLIP framework, and experimental results indicate that CLIP outperforms existing indoor positioning methods and can achieve centimeter-level accuracy with the improvement ratio of 94.6% over Sextant.

In this article, we propose a time-dependent lane-level navigation (TDLN) framework with spatiotemporal mobility modeling based on the Internet of Vehicles (IoV). The proposed TDLN framework can provide drivers with the fastest navigation path that can avoid passing congestion areas and predict vehicle spatiotemporal mobility of future traffic flows by estimating the travel time of road segments and the waiting time of intersections. According to our review of relevant research, TDLN is the first lane-level navigation solution that can provide the following features: 1) it can navigate vehicles in a lane-level manner and classify the queuing state of each vehicle as passing through an intersection; 2) it can estimate the driving time of lanes and the stopping time of intersections in different lanes to calculate the total delay time of passing through each lane and intersection; and 3) it can predict future traffic flows to determine the congestion level of each lane and explore predicted flow conditions on the road network to achieve the fastest navigation path planning. Simulation results show that TDLN outperforms existing methods and can plan the lane-level navigation path with the shortest travel time.

This paper proposes a time-dependent visiting trip planning (TVTP) framework to find the most efficient visiting order and plan the fastest moving paths based on Internet of Things (IoT) localization. The proposed TVTP framework consists of a deep learning based crowd density prediction model and a time-dependent visiting trip planning algorithm. In the developed prediction model, densely connected convolutional networks are explored with spatiotemporal data fusion to further reduce prediction errors. In the designed planning algorithm, visitors are guided to multiple target places at feasible time points to minimize total moving time based on predicted future crowd densities. According to our review of relevant research, this is the first framework that integrates deep learning for crowd density prediction with time-dependent planning for a multi-target visiting trip, which can precisely estimate future moving times based on predicted crowd densities and efficiently plan the optimal visiting order and guiding paths with the shortest total moving time to visit all target places. The open crowd dataset is adopted to evaluate the performance of existing works and TVTP. Experimental results show that our framework outperforms existing methods and can accurately predict the future crowd density of indoor people as well as significantly reduce the total moving time in the planned multi-target visiting trip.

This paper designs and implements an adaptive precise indoor positioning system, called Anchor-Few, for low anchor densities through Internet of Things (IoT) localization. Anchor Few exploits the IoT localization device, iBeacon, to provide accurate indoor positioning using only one or two anchors as three anchors are unavailable for triangulation. To the best of our knowledge, Anchor-Few is the first system that can provide the following features. First, the localization environments can be automatically classified based on the number of different iBeacon signals received by mobile devices, current moving direction, and historical location information. Second, it can use the distances from iBeacon nodes and previous user locations to detect whether the user is moving or not and determine the current moving direction. Third, in the environment covered by only one iBeacon node, a more accurate position than that using existing methods can be derived through the iBeacon position, distance from the iBeacon, and current moving direction. Finally, for environments covered by two iBeacon nodes, multiple iBeacon information and the user's historical locations can be cooperatively used for accurate position estimation. An Android-based prototype with deployed iBeacon nodes is implemented to verify the feasibility and correctness of our Anchor-Few system, which can achieve high localization accuracy while keeping anchor densities low.

In this article, we propose a mobile crowdsourced guiding and finding (MCGF) framework using smartphones to guide indoor people and find missing targets through Internet-of-Things (IoT) localization. The MCGF framework can cooperatively find lost/stolen targets equipped with mobile iBeacon nodes through participatory sensing networks formed by mobile users using smartphones in places with static iBeacon nodes. To precisely localize the missing target, fundamental target localization cases in distinct crowdsourced environments are formally classified and efficiently addressed to reduce the positioning errors with different numbers of smartphones detecting the missing target and different numbers of fixed iBeacon nodes nearby these target-detecting smartphones. According to our review of relevant research, this is the first solution that can provide the crowdsourced guiding path to a missing target with high localization accuracy for all densities of participating smartphones and iBeacon nodes. In particular, an Android-based prototype with static and mobile iBeacon nodes is implemented to verify the feasibility and superiority of our framework. Experimental results show that MCGF outperforms the existing methods and can significantly reduce the localization errors of mobile users and missing targets.

In this article, we propose a cooperative fleet cruise control (CFCC) framework for collision avoidance and lane changing based on vehicular sensor networks. The proposed CFCC framework is designed for fleets to reduce the recovery time to get the intended speed back after sudden braking, which can improve the driving efficiency of fleets. In addition, CFCC can determine the critical safety spaces of car following and lane changing for fleet vehicles to avoid the accident of rear-end and side collisions due to emergency braking. Furthermore, fleet lane changing can be coordinated between fleet vehicles in the original lane and neighboring vehicles in the target lane for ensuring safety and improving efficiency. According to our review of relevant research, CFCC is the first fleet cruise control system to provide the following features: 1) extending the car following risk of individual vehicles to the fleet following risk of fleet vehicles; 2) minimizing the acceleration and deceleration of fleet vehicles to reduce the speed recovery time; and 3) coordinating individual cars and fleet vehicles to provide the sufficient lane changing space as soon as possible. Simulation results show that our framework outperforms existing car following and lane changing methods and can significantly improve the driving safety and recovery efficiency of fleets as emergency events occur.

In this paper, we design and implement a deep advertising signage system, called DeepAd, with context-aware advertisement and cyber-physical interaction based on Internet of Things (IoT) technologies. In the DeepAd system, instant sensing and diverse interacting features are integrated with an IoT signage, which can (1) transmit multimedia contents and receive specific messages to/from smartphone users, (2) sense and interact with nearby individuals through image sensors, (3) embed context-aware advertisement information in sound waves, and (4) customize the on-screen 3D doll with an audience's face on demand. Through built-in sensors and smartphone interfaces, DeepAd can interact with nearby audiences via real-time multimedia services on the IoT signage in a click-and-drag manner. In addition, DeepAd investigates data-over-sound techniques to send embedded status-related advertisement via background music/voice. Furthermore, DeepAd explores deep learning based face changing and recognition to provide innovative and customized services to smartphone users. This paper demonstrates our current prototype consisting of the Android App, advertising server, and IoT signage.

In this paper, we design a driver behavior monitoring and warning (DBMW) framework to detect dangerous driving for enhancing road safety through the Internet of Vehicles (IoV). The designed DBMW framework applies onboard image sensors and wearable devices to detect the deviation degree of vehicles and trace the head motion of drivers, respectively. According to our review of relevant research, DBMW is the first framework for driver behavior monitoring and warning that provides the following features: 1) DBMW can keep recognizing the located lane lines and estimating the power spectral density of lane deviation for a vehicle through image sensors, 2) DBMW can keep monitoring driver behaviors and measuring the anomaly level of a driver through wearable devices, and 3) DBMW can instantly send the warning messages of potential dangerous driving to neighboring vehicles and nearby pedestrians through IoV communications as necessary, which makes vehicles and pedestrians be aware of the existence of surrounding dangerous drivers in advance to keep alerting and avoid potential accidents/collisions. In particular, the prototype consisting of an Android-based sensing unit and an Arduino-based wearable device is implemented to verify the feasibility and superiority of DBMW. Experimental results show that DBMW outperforms existing methods and can significantly improve the detection accuracy and false alarm rates of dangerous driving behaviors.

In this paper, we propose a time-efficient indoor navigation and evacuation (TINE) framework to minimize moving time for mobile users based on Internet of Things (IoT) technologies. In normal time, the proposed TINE framework can estimate the density of mobile users in each area and determine the moving speeds to pass through different areas. Based on the determined moving speed of each area, an indoor navigation path can be planned to provide the shortest moving time for a mobile user. In emergent time, TINE can accurately estimate the escaping time for groups of mobile users by jointly considering the length and moving time of passageways, the capacity of passageways/doors/exits, the present distribution and parallel moving of mobile users, and the possible congestion caused by other groups. Based on the estimated escaping time, TINE can efficiently alleviate the congestion of all passageways/exits and evenly distribute the evacuation load among exits to minimize the total escaping time. According to our review of relevant research, this is the first solution that can both provide the fastest navigation path to arbitrary target places and evacuate all groups of mobile users to safe places in the shortest escaping time. Simulation results show that TINE outperforms existing schemes and can significantly reduce the total walking and escaping times of indoor navigation and evacuation, respectively. In particular, an Android-based prototype with iBeacon IoT localization is implemented to verify the feasibility of our TINE system.

This paper proposes a smartphone-based cyber-physical system, called All-You-Can-Bike, for bikers to ride a bicycle around the world through street view and virtual reality based on Internet of Things (IoT) technologies. On the bicycle side, a microcontroller, a rotation sensor, and a Bluetooth module are used to calculate the cycling speed of the biker, sense the rotation period of wheels, and communicate with the biker's smartphone, respectively. On the biker side, a smartphone is used to detect the riding direction (through the built-in gyroscope), receive the cycling speed data (through Bluetooth communications), and determine the feasible frame updating rate of street view/virtual reality (through the proposed fuzzy control mechanism). Based on the current and historical cycling speeds, All-You-Can-Bike can derive the feasible frame updating rate based on the defined fuzzy sets of wheel rotation periods and cycling acceleration. An Android-based prototype with the Arduino developing board is implemented to verify the feasibility and superiority of our All-You-Can-Bike system.

This paper presents a centimeter-grade metropolitan positioning (CGMP) framework for achieving lane-level localization through Internet of Vehicles (IoV) communications. The CGMP framework integrates vehicle-to-roadside (V2R) and vehicle-to-vehicle (V2V) communications with installed roadside/intersection cameras to provide positioning information to vehicles on the road. This framework can be used to estimate the precise position of a vehicle and detect the lane of each vehicle. The estimated positions and detected lanes can be broadcast to nearby vehicles through V2R and V2V communications. According to our review of relevant research, CGMP is the first positioning system to provide the following features: centimeter-grade positioning accuracy through IoV communications that can detect a vehicle's current lane; the employment of existing roadside/intersection cameras, thus keeping additional construction costs low; and an innovative infrastructure for the application of next-generation intelligent transportation systems that can enable lane-level traffic control, collision avoidance, and vehicle navigation. A prototype based on wireless access in vehicular environments/dedicated short-range communications is implemented to verify the feasibility and effectiveness of the proposed framework. The experimental results indicate that fully deployed CGMP can achieve centimeter-grade positioning accuracy with the error around 10 cm, and partially deployed CGMP can significantly reduce the positioning errors of existing localization methods.

This paper proposes a smartphone-based indoor navigation and evacuation system, called EasyGO, to minimize moving time for mobile users through Internet of Things (IoT) technologies. In normal time, EasyGO can estimate the density of indoor people in each area and determine the moving speeds to pass through different areas. Based on the determined moving speed of each area, an indoor navigation path can be planned to provide the shortest moving time for EasyGO users. In emergent time, EasyGO can accurately estimate the escaping time for each EasyGO user by considering the capacity and length of passageways, the capacity of exits, and the current distribution and parallel moving of indoor people. Based on the estimated escaping time, EasyGO can evenly distribute the evacuation load among exits and alleviate the congestion of all passageways and exits to minimize the total escaping time. An Android-based prototype with iBeacon indoor localization is implemented to verify the feasibility of our EasyGO system.

In this paper, we design an audience-aware signage sensing and interacting system, called Cyber-Physical Ad (CPAd), for individuals using smartphones based on Internet of Things technologies. In the CPAd system, a digital signage can sense surrounding audiences via built-in sensors and the signage content can be adaptive to interact with nearby audiences through real-time multimedia contents. In addition, audiences can use smartphones to click and drag a dedicated service on the touchscreen to an endorser face on the digital signage. An Android-based prototype of CPAd is implemented to sense approaching audiences and interact with these audiences in a cyber-physical manner to achieve effective advertising and marketing.

This paper proposes a group-based framework with dedicated path planning for emergency guiding based on Internet of Things (IoT) technologies. The proposed framework can model the spatiotemporal mobility of indoor people to determine and relieve the congestion of corridors and exits. A dedicated path can be determined to provide the shortest evacuation time for each group of nearby people. The corridor and exit capacities, corridor lengths, clustering motion of a group, concurrent moving of different groups, and up-to-date distribution of group people are considered together to accurately estimate the evacuation time for each group. Based on the estimated evacuation time, evacuation load can be evenly distributed among corridors and exits to alleviate the congestion of all corridors and exits for minimizing total evacuation time. The performance of the proposed framework is evaluated by conducting mathematical analysis and computer simulations, which outperforms existing schemes and can achieve the shortest evacuation time for group-based emergency guiding. In addition, an Android-based prototype with indoor IoT localization technologies is implemented to verify the feasibility of our framework.

Traffic congestion is a critical concern in most cities. Inefficient traffic control wastes time and fuel, and causes harmful carbon emissions, road accidents, and many economic problems. This paper proposes a cooperative traffic control framework for optimizing the global throughput and travel time for multiple intersections. Adjacent intersections are considered in analyzing their joint passing rates and attempting to maximize the number of vehicles traveling through a road network. The proposed framework provides fairness for each road segment and realizes the green wave concept for arterial roads. This paper extends previous studies by considering the passing rates of continuous road segments and coordinating traffic signals of multiple intersections. The simulation results show that the approach outperforms existing schemes in that it achieves a high global throughput, reduces the average waiting time, lowers the total travel time, and decreases average CO2emissions. To verify the feasibility of the proposed framework, a wireless access in vehicular environments/dedicated short-range communications-based prototype for lane-level dynamic traffic control is designed and implemented.

This paper proposes a load-balancing framework for distributed emergency guiding based on wireless sensor networks. A load-balancing guiding scheme is designed and an analytical model is derived to reduce the total evacuation time of people indoors. The guiding scheme can provide the fastest path for people to reach an exit according to the evacuation time estimated using the analytical model. Based on thorough research, this is the first distributed solution in which corridor capacity and length, exit capacity, and the concurrent movement and distribution of people are considered in estimating the evacuation time and planning escape paths. Using the proposed framework, congestion in corridors and at exits can be eased to substantially reduce the total evacuation time. Analytical and simulation results show that this approach outperforms existing schemes and can prevent people from following localoptimal guiding directions that increase the evacuation time. A prototype called the Load-balancing Emergency Guiding System (LEGS) is implemented; this system can be used to compare the evacuation times and guiding directions provided by existing schemes and the proposed scheme for various distributions of people.

This paper proposes a lane-level beacon-less, infrastructure-less, and GPS-less cooperative collision avoidance (BIG-CCA) framework for preventing rear-end collisions. The BIG-CCA framework applies vehicular sensor networks to prevent chain vehicle collisions, which are common road accidents that occur when vehicles make sudden stops. BIG-CCA enables vehicles equipped with only onboard sensors to prevent such accidents. Based on extensive research, BIG-CCA is the first lane-level CCA solution that provides the following features: 1) BIG-CCA does not maintain a list of neighboring vehicles through beacons and, thus, the overall signaling overhead can be reduced to conserve bandwidth; 2) BIG-CCA does not use the GPS positions of vehicles to facilitate collision avoidance and, thus, the inaccuracy and unavailability of a GPS can be avoided; and 3) BIG-CCA does not rely on costly roadside infrastructures but employs only vehicle-to-vehicle communications to form warning groups for vehicles driving along in the same lane. BIG-CCA consists of a distributed grouping mechanism and a receiver-based forwarding scheme. Vehicles join or leave a warning group through only single-hop transmissions with the low overhead of group maintenance. When sudden braking occurs in a warning group, the proposed receiver-based forwarding scheme can specify a single forwarder without message contention. The performance of BIG-CCA is evaluated by conducting mathematical analysis and computer simulations, which outperform existing methods. An Android-based prototype is also implemented to verify the feasibility of BIG-CCA.

This paper proposes a group-based emergency guiding system for mobile users using smartphones, called Go FAST, which can model the spatiotemporal mobility of indoor people. A dedicated path can be determined to provide the shortest evacuation time for each group of nearby people. The Go FAST system considers the corridor capacities and lengths, exit capacities, concurrent motion, distribution of indoor people, and dedicated escape path to accurately estimate the evacuation time for each group. Based on the estimated evacuation time, the evacuation load can evenly distributed among exits to minimize the total evacuation time. Go FAST can alleviate the congestion of all corridors and exits to reduce the total evacuation time as much as possible. An Android-based prototype with iBeacon indoor localization is implemented to verify the feasibility of our Go FAST system. Simulation results show that Go FAST outperforms existing schemes and can achieve the shortest evacuation time for group-based emergency guiding.

In this paper, we design an augmented reality based social networking system, called SocialYou, for mobile users using smartphones. The SocialYou system only needs a single-finger action that drags the icon of social networking websites to the target face on the screen of a smartphone. The target user webpage in a specific social networking website can be immediately displayed in the social media App or built-in web browser. Through SocialYou, mobile users do not need to know the target identifier in advance. In addition, mobile users do not need to input identifier information by themselves. Furthermore, mobile users can select a specific target among multiple candidates from the camera of a smartphone. SocialYou reveals an innovative user interface for social networking between mobile users. We have implemented an Android-based SocialYou system that outperforms existing methods in experimental results.

This paper proposes an individual-based framework for emergency guiding. The spatial-temporal mobility of all people is modeled to determine a dedicated path that provides the shortest evacuation time for each person. According to our review of relevant research, this is the first optimal solution without using time-expanded graphs, and corridor capacities and lengths, exit capacities, concurrent motion, and distribution of people are considered to minimize evacuation time. We prove that the proposed path planning algorithm is optimal and analyze its time and space complexity. The proposed framework can be used to estimate the evacuation time for each person accurately and evenly distribute evacuation load among exits to achieve the most efficient load balance. In the proposed framework, the congestion in all corridors and exits can be alleviated to maximally reduce the total evacuation time. Simulation results show that our approach outperforms existing schemes, and can be used to determine an optimal escape path for each person and, thus, achieve the shortest total evacuation time.

In this paper, we design an augmented reality based file transfer system called ShareAR. The ShareAR system only needs a single-finger action that drags the file icon to the target face on the screen of a mobile device. The dragged file can be transmitted to the target receiver who gets the file immediately as her/his mobile device is on-line or receives the file later after connecting to the Internet. Through ShareAR, users do not need to know the target account in advance. In addition, users do not need to input account information by themselves. Furthermore, users can select a specific target among multiple candidates from the camera of a mobile device. ShareAR reveals an innovative user interface to transfer files between mobile devices. We have implemented an Android-based ShareAR system that outperforms existing schemes in experimental results.

In this paper, we design and implement a lane-level cooperative collision avoidance (LCCA) system using vehicle-to-vehicle communications. The LCCA system applies vehicular sensor networks to preventing chain vehicle collisions, which allows vehicles with merely onboard sensors to prevent such collisions on the road because of sharp stops. To the best of our knowledge, this is the first CCA system that does not use inaccurate GPS locations and costly roadside infrastructures to avoid chain vehicle collisions. LCCA employs inter-vehicle communications and onboard sensing to form warning groups, where each warning group is a set of vehicles that drive along the same lane and every pair of adjacent cars is within a certain distance. Only single-hop transmissions are needed to join/leave a warning group, thus keeping the group maintenance overhead low. When a sudden braking is taken in a warning group, LCCA can quickly propagate warning messages among group members. This paper demonstrates our current prototype.

Traffic congestion in modern cities seriously affects our living quality and environments. Inefficient traffic management leads to fuel wastage in volume of billion gallons per year. In this paper, we propose a dynamic traffic control framework using vehicular communications and fine-grained information, such as turning intentions and lane positions of vehicles, to maximize traffic flows and provide fairness among traffic flows. With vehicular communications, the traffic controller at an intersection can collect all fine-grained information before vehicles pass the intersection. Our proposed signal scheduling algorithm considers the flows at all lanes, allocates more durations of green signs to those flows with higher passing rates, and also gives turns to those with lower passing rates for fairness provision. Simulation results show that the proposed framework outperforms existing works by significantly increasing the number of vehicles passing an intersection while keeping average waiting time low for vehicles on non-arterial roads. In addition, we discuss our implementation of an Zigbee-based prototype and experiences.

The Eco-Sign system is a traffic light control system for minimizing greenhouse gases emitted by idling vehicles at intersections. Eco-Sign provides the following features: (i) it can notify vehicles to turn on/off their engines based on expected waiting time for green lights at intersections, (ii) it can dynamically adjust traffic light timing to minimize the number of vehicles stopping at an intersection based on vehicle arrival and departure rates, and (iii) it is a fully distributed system in the sense that each intersection can learn its local traffic condition and optimize its traffic sign setting to prevent congestions and thus traffic jams. Eco-Sign thus demonstrates a new traffic light control system for environmental protection.

In this paper, we propose a Vehicular Surveillance and Sensing System (VS3), which targets at car security and tracking applications. VS3 can be triggered by events detected inside or outside a car, such as abnormal air quality, potential burglary, and identification of some target vehicles (such as stolen cars). Via a 3G module, a user can interact with VS3 via multimedia communications. For security applications, we show how VS3 detects an abnormal CO2 level or potential car burglary, notifies the vehicle owner, and then interacts with the owner. For tracking applications, we show how VS3 identifies potential stolen vehicles, transmits reports to the police department, and get neighboring cars involved to cooperatively track suspicious vehicles. This paper demonstrates our current prototype.

The Vehicular Surveillance and Sensing System (VS3) is a 3G-based mobile device for car security applications. On the car side, it consists of a CO2sensor, a camera module, a 3G module, and a microprocessor. On the user side, only a 3G mobile phone is needed. VS3provides the following features: (i) it can be triggered by events detected on car, (ii) events can be abnormal air quality or potential burglary, and (iii) it supports text or multimedia interaction with users. Application scenarios include detecting an abnormal CO2level or potential car burglary, which triggers VS3to transmit SMS, MMS, or interactive video call to the vehicle owner, who can then monitor the car situation in return. VS3thus demonstrates a new car security and burglarproof prototype.

The IEEE 802.16 standard for wireless metropolitan area networks (WMAN) has been created to meet the need of wide-range broadband wireless access at low cost. The objective of this paper is to study how to exploit spectral reuse in an IEEE 802.16 mesh network through timeslot allocation, bandwidth adaptation, hierarchical scheduling, and routing. To the best of our knowledge, this is the first work which formally quantifies spectral reuse in IEEE 802.16 mesh networks and which exploits spectral efficiency under an integrated framework. Simulation results show that the proposed spectral reuse scheduling and load-aware routing significantly enhance the network throughput performance in IEEE 802.16 mesh networks.

The mobile ad hoc networks (MANETs) have received a lot of attention recently. While many routing protocols have been proposed for MANETs based on different criteria, few have considered the impact of multi-rate communication capability that is supported by many current WLAN products. Given a routing path, this paper provides an analytic tool to evaluate the expected throughput of the route, assuming that hosts move following the discrete-time, random-walk model. The derived result can be added as another metric for route selection. Simulation results are also presented.

This paper considers the location tracking problem in PCS networks. How a solution to this problem performs in fact highly, depends on the mobility patterns of users. In this paper we propose a new traveling salesman mobility (TSM) model, in the hope of catching the mobility patterns of a large group of users. The TSM model is characterized by features of "stop-or-move", "infrequent transition" "memory of roaming direction", and "oblivious in different moves". Then a location tracking strategy based on this TSM model is developed. The scheme only needs to keep very little information for each user. Analyses and simulations are provided, which show that the strategy is very prospective.