FishEye Matlock: A Random Functional Encoding Mechanism for Secure Location Sharing

Abstract

Location tracking is difficult to protect due to the sequential nature of the data and the need for accuracy to offer a proper service and monetize the location information. Homomorphic encryption can partially solve the problem, but it typically has a minimal set of operations that are not feasible for geographical operations. Randomized Functional Encoding (RFE) allows the creation of random keys to decrypt data, according to the user's needs. Now, creating keys that only focus on a portion of the path while protecting the rest of the path has not been proposed in the literature, to the knowledge of the authors. This work proposes a RFE mechanism for Matlock-coded location data, called FishEye Matlock. This technique generates disposable random key matrices that only reveal a desired portion of the path, with the possibility for the user to add random noise to protect the revealed data and to control the amount of noise added to the rest of the path. This allows secure information sharing with particular actors, like law enforcement, so that the information of interest is shared without affecting the user's privacy. The algorithm is tested in two different path scenarios to show the technique's applicability, the level of protection, and the impact of the parameter value selection. Results show that the mechanism can be tailored to generate key matrices for different scenarios: at the lowest value of k, the level of noise reaches several thousands of kilometers of noise along the path, and between 60 and 100 times the level of noise, and with, the highest k value, between 40% to 80% of the maximum distance radius on average at the point of interest, and between 1.1 and 10 times the defined noise level at the path.