Remote Identification supports unmanned aerial vehicle accountability, but routine broadcasts are openly observable and copyable. The selected post-quantum key-encapsulation and signature objects exceed the compact Broadcast Remote ID authentication capacity, while per-pseudonym credentials impose linear provisioning and storage costs. This paper presents a commitment-based architecture in which a Trusted-Authority-signed Merkle root authorizes
n interval pseudonyms. Routine broadcasts carry a 128-bit pseudonym; an authorized verifier obtains an ordered Merkle path and transcript signature through a unicast exchange protected by post-quantum key encapsulation and authenticated encryption. Authority provisioning and nonvolatile credential state are
O(1) per root, whereas tree construction, activation, and volatile state remain
O(n) . A modeled ring-oscillator physically unclonable function is evaluated for credential reconstruction using top-950 selection and shortened Bose–Chaudhuri–Hocquenghem decoding; Toeplitz extraction is benchmarked separately. Across 1300 pre-silicon runs, 500 nominal virtual devices produced approximately 50% uniformity and inter-instance Hamming distance. Of 100 fixed-mismatch PVT devices, 54 enrolled and all 54 reconstructed successfully across eight nonnominal corners. Processor evaluation showed about 4.5 ms Raspberry Pi 5 authentication latency and
5.64±0.14 mJ per authentication at
n=16384 . ESP32 processor measurements demonstrated execution of the selected post-quantum operations; a UAV-mounted ESP32 completed 120/120 recorded outdoor-hover authentication exchanges using provisioned credentials, without physical PUF reconstruction. Controlled mixed-load and network-impairment tests incurred no legitimate-request failures; Tamarin verified authentication and revocation properties. Verifier acceptance establishes credential possession and authorized-set membership; it does not provide physical-source verification or relay resistance.
In this paper, we propose a leakage-resilient and proactive authenticated key exchange (called LRP-AKE) protocol for credential services which provides not only a higher level of security against leakage of stored secrets but also secrecy of private key with respect to the involving server. And we show that the LRP-AKE protocol is provably secure in the random oracle model with the reduction to the computational Difie-Hellman problem. In addition, we discuss about some possible applications of the LRP-AKE protocol.
Abstract. Nowadays, most of sensitive applications over insecure network are protected by some authenticated secure channel which is highly relies on specific authenticated key exchange (AKE) protocol. Nevertheless, the leakage of authentication credential used in AKE protocol somehow result in unauthorized exploitation of credential information via identity impersonation (IDI) attack. To address the problem of IDI, we introduce a new dynamic authentication factor for AKE protocols, i.e., the secret execution states, to either prevent IDI attack by detecting attempts thereof, or limit its consequences by on-line detecting situations of previously unidentified IDI. In this paper, we model the security for authenticated key exchange with synchronized states (AKESS) based on Bellare-Rogaway model, and we particularly formalize the IDI and IDI detection. We propose a generic execution states synchronization framework for AKE, in which we utilize the session key to generate the secret execution states on both sides, and present a new AKESS protocol which is provably secure in the standard model. Our goal is to enhance the security of existing authenticated key exchange with long-lived key (AKELL) protocols by equipping them with the capabilities of both IDI prevention and detection without modifications on those protocols.