Existing anonymisation schemes often rely on static identifiers, which are vulnerable to long-term observation and incur high overhead due to public key operations, while offering only a single anonymity level. We propose a personalized, secure, anonymous traceability mechanism with pseudo-random TTLs inside a trusted execution environment (TEE). The scheme provides three adjustable anonymity levels and resists long-term observation attacks. Formal analysis proves its security and anonymity. Compared to the existing state-of-the-art anonymisation scheme EMMCRS, experiments show that our design reduces computational overheads by 99.3%, 98.8%, and 98.8%, and reduces communication overheads by 74.5%, 62.2%, and 50.3%.