In this paper, we study the problem of accurate tracking of a mobile target by a central authority, using distance estimates obtained by a group of untrusted anchors within the communication range of the target. We show how to perform accurate localization of the target in the presence of some compromised and colluding malicious anchors that lie about the position of the target. We also show how to identify most of these malicious anchors. In the case where measurements are error-free, we derive an upper bound (B) on the number of malicious anchors that may be involved in localizing the target while still not being able to undermine its accurate localization. We propose a scheme to correctly localize the target when the number of malicious anchors within its range is no more than 13. It also identifies all the malicious anchors. In the presence of positive measurement errors, we propose a scheme based on convex optimization that can localize the target despite the presence of an arbitrary number of malicious anchors in its range. When the number of malicious anchors are no more than B, our scheme localizes the target with an error less than 1m and is also able to identify more than 80% of the malicious anchors. Both our schemes are simple and easy to implement.