Abstract
The rapid proliferation of Internet of Things (IoT) devices in resource-limited environments poses significant security challenges, as traditional cryptographic methods are often too slow and resource-intensive for systems with limited power and memory. This study addresses these issues through the development of a novel hybrid cryptosystem that combines Deterministic Finite Automata (DFA)-based hashing with Lorenz attractor-driven encryption for secure IoT data transmission. The approach is twofold: first, hash values are generated from fixed sensor data sources; second, the Lorenz chaotic system uses this data as initial conditions to generate unpredictable encryption keys. The system employs lightweight XOR-based encryption, making it well suited for microcontrollers. A working prototype was developed and implemented on Arduino platforms using the Wokwi simulation environment and tested with real sensor inputs (temperature and photoresistor). Experimental results show excellent performance: zero hash collisions for 1,000 input values ranging from 0 to 1023, total key unpredictability even with minimal input changes, and encryption time under 120 ms on Arduino Uno. The system achieves robust security and 100% accuracy in tampering detection, while consuming less than 20% of the available memory on the microcontroller. Its successful implementation demonstrates the feasibility of hybrid DFA-chaos cryptography in embedded IoT environments.