Po2V: Network Layer Position Verification in Multi-Hopo Wireless Networks

The correctness of the position is fundamental for position-based services. Previously proposed position verification schemes require an infrastructure, or the existence of a one-hop direct communication between the prover and the verifier. These schemes are less feasible in infrastructure-less networks where communication is achieved via multiple hops, such as mobile ad hoc networks. In this work we propose PO'V, a lightweight, network layer position verification scheme for distributed position services in mobile ad hoc networks, under which a position server can verify whether a user has reported a correct position. We use adaptive transmitting power and multi-path polling to improve verification accuracy. We study different attacks against the verification scheme and propose corresponding mitigation techniques. \I1e use a reputation system to reduce the false positives in a network with and without attackers. Our analysis and simulation studies show that while less expensive because no advanced techniques are required in the physical layer, PO'V can achieve a verification accuracy that suffices for many applications. Location-based services (LBS) [I]: [Z] have become an important part of pervasive computing services. A LBS server uses the location of a user to determine whether the user should be served. Location or position information is also used in mobile ad hoc networks to facilitate routing. Position-based (or geographic) routing protocols [3] make decisions based on the geographical position of the destination of a packet, and therefore have better scalability and routing performance (e.g., delivery ratio, end-to-end delay) than traditional routing protocols. In sensor networks, the location of a sensor determines whether the information collected by the sensor is useful, as the location indicates what area the sensor is monitoring. The correctness of position is essential for position-related services. For example. location-based access control allows a client to obtain access to a service based on location. A malicious user may claim to be within an area that gives him fraudulent access to a service. In the context of sensor networks, many sensor applications require the knowledge of the origin of the sensed information. A malicious participant can pretend to be at critical positions and provide misleading information. In addition. a false position in sensor networks facilitates attacks such as sensor displacenlent (a sensor is temporarily moved out of the network by an attacker with the goal to compron~ise itt but the sink will believe that the sensor is still in the network), wormhole attack (a sensor thinks a node far away from i t is its neighbor) and false network topology (a base station makes wrong routing decisions based on false sensor positions). In ad hoc networks: as positionbased routing relies on correct positions. a false position of the destination will result in a routing failure. Moreover. an attacker can make a neighbor to believe that the attacker is the closest to the destination and be selected on the routing path: by manipulating the position infonnation. As a result. the attacker will obtain control of significant traffic in the network. Finally. the correctness of the position information is essential for secure protocols that use position to prevent certain threats against network services [4], [5]. Position services can be classified in i1erwo1-k-ceilrric if the network obtains the position of the device directly. or device-ceilrric if the method of obtaining the position relies on a device (i.e. a mobile LBS user. an ad hoc node, or a sensor node). For example: in a network-centric approach. the position of a user is leanled based 011 the access point he is connected to or the base station he-obtains service from, since he must be within their coverage area. More accurate network-centric position obtaining approaches can be found in [6], [7]. In comparison, in a device-centric service, the position of a device is obtained by the device itself: e.g., according to the navigation signals from GPS [8] satellites or the beaconing signals from pre-deployed landmarkers [9]. The position is reported to so-called position servers, which store the positions. Other network users can retrieve the position information from the position server for different network functions. Device-centric position service systems have been proposed in mobile ad hoc networks in [I O]? [I I]. [I 21. [ I 31. An attacker can generate false position information and interrupt network-centric position-based services by disrupting the position calculation function. In the case of device-centric position-based services: an attacker can sin~ply report a false position or attack the navigation signal in GPS or the beaconing signal in sensor networks, ensuring that the positions obtained by honest users are not correct. Position verification is a critical service. Its absence can result in numerous problems as discussed above. In the rest of the paper, we call the device of which the position needs to be verified a prover, and the parties that verify the position ~*er i ' ers . Previously proposed position verification schemes [15], [14]. [16], [I71 require an infrastructure. because the 0 2V: I