# Rotary Electromechanical System Integrating Non‐Reciprocal Memory and Combinational Logic

## TL;DR

This paper introduces a rotary electromechanical system that combines memory and logic for reprogrammable decision-making, advancing physically embedded intelligence.

## Contribution

The novel integration of non-reciprocal mechanical memory with combinational logic in a single rotary system enables reprogrammable sequential computation.

## Key findings

- The system functions as a four-bit finite-state machine with reconfigurable state-transition rules.
- It demonstrates functionalities like digital combination locking and in-memory electromechanical computation.
- Serial coupling of bistable units creates a history-dependent, non-reciprocal mechanical memory.

## Abstract

Recent advances in mechanical computing have harnessed bistable mechanisms with intrinsic memory to extend the scope of physically embodied intelligence, enabling history‐dependent behavior. However, existing mechanical computing architectures largely fail to integrate mechanically encoded memory with the logic operations necessary for sequential information processing, a foundational requirement for advanced computational tasks, such as autonomous sequential decision‐making. Here, we introduce a rotary electromechanical computing system that unifies non‐reciprocal mechanical memory with combinational logic, enabling reprogrammable sequential decision‐making within a single finite‐state‐machine (FSM) framework. The architecture consists of serially coupled rotary bistable units that collectively produce a history‐dependent, non‐reciprocal mechanical memory. The discrete geometric orientation of each unit encodes a binary state, which is transduced through a conductive network to execute logic operations. Under applied torque, the stacked system functions as a four‐bit FSM, in which state‐transition rules can be reconfigured by modifying the electromechanical coupling. This rotary FSM demonstrates multiple computing functionalities, including digital combination locking, in‐memory electromechanical computation, and reprogrammable digital control. By embedding non‐reciprocal state evolution and electrical logic directly into physical hardware, this work establishes a pathway toward physically embedded intelligence for next‐generation electromechanical systems.

Rotational bistability, interfaced through a conductive network with switchable contacts, unifies logic operations and mechanical memory within a rotary electromechanical system. The author demonstrates how serially coupled configurations of rotationally bistable modules enable reprogrammable sequential logic governed by mechanical state transition and electrical connections, a step toward physically embedded intelligence.