Ranjeet Singh Tomar

Prediction of Lane Change Trajectories through Neural Network

Lane changing process is an essential maneuver, however, the process is responsible for large number of collisions and traffic instability. In this work, the effectiveness of neural network for prediction of future lane change trajectory based only on the past vehicle path is presented. Existing lane change models and lane change process do not consider the uncertainties and perceptions in the human behavior that are involved in lane changing. A neural network may learn and incorporate these uncertainties to predict the lane changing trajectory in the near future more accurately. A multilayer perceptron (MLP) has been employed to train itself from existing NGSIM field data and predict the future path of a lane changing vehicle. The impact and effectiveness of the proposed technique is demonstrated. Prediction results show that an MLP is able to give the future path accurately only for discrete patches of the trajectory and not over the complete trajectory. The results confirm to the observation that a vehicle trajectory has immediate influence from its neighborhood whose information is imperative for trajectory prediction.

CMAC: A cluster based MAC protocol for VANETs

The messages in a VANET environment are short with limited time value. The aim of a MAC protocol is to deliver these messages with high reliability and low delay. To guarantee this in a high mobility and fluid topology environment, the contention period over the shared wireless channels must be minimized. In this work, an RSU centric cluster based MAC (CMAC) protocol has been proposed. In this protocol, the channel allocation and management is transferred to the RSU to obviate channel contention for reliable delivery of messages. The RSU region is divided into prefixed clusters with frequency reuse in non adjacent clusters. The frequency reuse increases the available bandwidth and reduces the waiting time of a node for channel allocation. The traffic and the VANET topology is generated separately and used to determine the effectiveness of CMAC protocol. Results indicate that the protocol has high throughput even when the number of vehicles is high and also for large clusters in an RSU region. This high packet reception ratio indicates the efficacy of the protocol for high reliability low delay delivery of messages in a VANET environment.

Suitability of MANET Routing Protocols for Vehicular Ad Hoc Networks

In this paper, the suitability of the existing MANET routing protocols for VANETs has been evaluated. Even though a VANET is a subset of MANET, it has unique characteristics and message delivery requirements. VANETs are characterized by a very dynamic topology with partial infrastructure support, patterned mobility, mobile nodes with sufficient resources, intermittent connectivity and varied channel behavior. In addition to these characteristics, the network traffic requirements are also different. Safety messages need reliable, accurate and timely delivery in a few hops neighborhood whereas other types of messages may have other requirements. MANET protocols, in contrast, are designed for resource constrained wireless environment. The message delivery constraints are not stringent and movement of nodes is slow and random. Thus, there is a need to ascertain whether, the basic MANET routing protocols like AODV and DSR can satisfy VANET needs like timely delivery of messages especially when the vehicles move at high speeds. Simulation studies show that AODV and DSR are not able to satisfy the requirements and specific VANET routing protocols are needed.

RSU centric channel allocation in Vehicular Ad-hoc Networks

Safety and non safety related VANET messages have localized spatial scope and limited time relevance. A MAC protocol must ensure timely delivery of such messages with high reliability over a shared wireless medium. The Vehicles moving on the road form a fluid strait jacketed topology. In this paper, an RSU centric channel allocation protocol has been proposed to minimize channel allocation time and management overhead. An RSU divides the limited bandwidth allocated to a region into prefixed overlapping spatial clusters and the channel in each cluster is divided into time slots. A time slot is allotted to a vehicle in accordance to the priority of the request and availability of the channel. Simulation results indicate that the allocation scheme is able to deliver packets within the stipulated time with high reliability for different vehicle speeds and density for channel allocation based on three overlapping clusters. Moreover, the technique does not suffer from the hidden and exposed node problem.

Safety of Lane Change Maneuver Through A Priori Prediction of Trajectory Using Neural Networks

Lane change (LC) is a maneuver that allows drivers to enter into a lane that suits their requirements and comfort. The LC process requires the driver to assess its neighborhood traffic in its original and target lanes before undertaking the maneuver. Other vehicles in the neighborhood also need to adjust for safe lane change. The LC trajectory is determined by the accuracy of these subjective assessments as well as the state of traffic. An erroneous assessment by LC vehicle or neighboring vehicles or an incorrect maneuver can cause collision. The collision can be prevented if the LC trajectory can be predicted and the feasibility of LC can be communicated to different vehicles involved in this process. . In the present paper, neural network is used for long term forecast of the lane change trajectory and for short term near future positions of the LC vehicle. The neural network is trained using past LC trajectories of different vehicles. The trained network is then used for long and short term forecast of the vehicle’s positions during LC. Simulation results with actual filed data observed data indicates that neural network is able to learn LC maneuvers and is able to perform short term prediction with sufficient accuracy.

Enhanced SDMA for VANET Communication

VANETs aim to improve traffic safety and driving experience on the road. A medium access protocol for VANETs determines their performance. A key challenge in the design of MAC protocol is timely and accurate delivery of messages even when network topology changes quickly. In this work, we propose enhancements in the native SDMA protocol to suit safety critical message transfer in VANETs. In the proposed enhanced SDMA, the road is logically divided into spatial slots and TDMA time slots are assigned to each spatial location. In Enhanced SDMA, the spatial slot allocation is done using the vehicular traffic distribution. The empty slots in sparse traffic and paucity of slots in congested traffic are allocated to optimize the fairness in channel distribution and throughput maximization using SINR (Signal to Noise and Interference Ratio). Simulation results indicate the Enhanced SDMA is able to achieve both efficient slot allocation and enhanced throughput for both free flow and congested traffic states.

Cluster Based RSU Centric Channel Access for VANETs

Safety related VANET messages have localized spatial scope and limited time relevance. A MAC protocol should exploit the unique VANET characteristics like restricted mobility in one dimensional motion paths, presence of RSU, limited transmission range of vehicles and large transmission range of RSU to ensure timely and reliable delivery of messages. In this paper, an RSU assisted TDMA protocol has been proposed to minimize channel allocation time and management overhead. The problems are the limited bandwidth, shared wireless medium, high mobility of vehicles etc. For this, the RSU divides the limited bandwidth allocated to its region into prefixed overlapping spatial clusters and the channel in each cluster is divided into time slots. A time slot is allotted to a vehicle in accordance to the priority of the request and availability of the channel. Simulation results indicate that the allocation scheme is able to deliver packets within the stipulated time with high reliability for different vehicle speeds and density for channel allocation based on three overlapping clusters. Moreover, TDMA based MAC with contention based request increases the packet reception ratio along with low in channel acquisition delay.

Trust Management Model for Wireless Ad Hoc Networks

Wireless ad hoc networks consist of a set of nodes equipped with wireless interfaces. The mobile / static nodes dynamically create a wireless network without using any infrastructure. In the absence of infrastructure authentication and trust management is an issue for nodes in wireless ad hoc networks. The proposed trust management model allows nodes to evaluate the trust by taking into account certificate of other node(s) to overcome vulnerabilities. In this paper, we have addressed the trust management using certificate based approach. Simulation results show that the time taken by the scheme is significantly less and transmission over the shared wireless channel is in the order of milliseconds for different traffic conditions in wireless ad hoc networks. This proves the efficacy of the scheme and makes it attractive for MANET and VANET.

Neural network based lane change trajectory prediction in autonomous vehicles

During a lane change, vehicle collision warning systems detect the likelihood of collision and time to collision to warn vehicles of an imminent collision. In autonomous systems, a vehicle utilizes data obtained by its own sensors to predict future state of traffic. The data from on board sensors are limited by line of sight, measurement noise and motion parameters of the vehicle which affect the accuracy of prediction. Alternatively, in cooperative driving, vehicles transmit their parameters continuously. This is also beset by communication delays and message loss. To avoid these limitations, a vehicle should estimate its future state and broadcast it to others vehicles in the neighborhood. This necessitates vehicles to predict their future trajectories based entirely on its past. Since, low cost global positioning systems are becoming an integral part of vehicles; a vehicle knows its own position. This can be utilized by the vehicle for prediction of its future trajectory. In this paper, the effectiveness of lane change trajectory prediction on the basis of past positions is studied. The lane change trajectory of a vehicle is modeled as a time series and back propagation neural network is trained using real field data and its efficacy for short range and long range prediction is benchmarked. Simulation results using NGSIM data indicate that future lane change trajectory cannot be predicted with sufficient accuracy. The most important reason seemed to be the influence of other neighboring vehicles on the trajectory on the lane changing vehicle in addition to noise and complex dependence of future on the past values. The results also indicate that a vehicle changes its motion parameters during the entire lane change process. This confirms the active intervention of the driver in adjusting the trajectory on the basis of his assessment of the future state of its surrounding vehicles and entails the consideration of the state of surrounding vehicle for accurate prediction.

Cryptographic Overhead in Sensor Networks with HFE Scheme

In this paper, the energy consumption of Hidden field equation (HFE) multivariate scheme in wireless sensor networks (WSNs) has been examined. Security provisioning in a WSN requires balancing between two conflicting goals - security strength and resource efficiency. Most of the cryptosystems with security strength required in a WSN are highly resource intensive. Cryptographic operations consume processor power, time and memory space for storage of intermediate and final output. The cryptographic bits also constitute communication overhead that induces delay and decrease in throughput. To lessen the load on the resource constrained WSN without compromising security, HFE multivariate scheme is proposed. It is a type of multivariate cryptosystems in which linear equations are solved to obtain successive sets of central variables. The small number of operations and limited memory requirements make this scheme a potential candidate for security provisioning in WSN. Analysis shows that the scheme can be implemented on resource constrained sensor nodes. And cryptographic overheads induced by HFE scheme are within the tolerable limits of WSN applications.