Journal of Advances in Mathematics and Computer Science

This paper proposes an adaptive PWM consensus strategy for multi-agent systems over undirected connected graphs. The communication topology is assumed to be an undirected connected graph. Each agent computes its signed distance to the midpoint of its neighbors’ convex hull. The control input is a fixed-amplitude pulse whose width is modulated by an adaptive gain. The gain increases for large errors to accelerate convergence and decays to a minimum when the agent is inside the convex hull, saving energy. Control stops once the agent is strictly inside the convex hull. Theoretical analysis guarantees state boundedness, gain boundedness, and global asymptotic consensus. Numerical simulations on a five-agent network show that the proposed method achieves perfect consensus with about 22.8 % energy saving compared to the algebraic-sum PWM, reduces the convergence time from approximately 10 s to 5 s, and attains zero consensus error at t=5.2 s while the algebraic-sum PWM still shows an error of 0.4. Fixed-amplitude pulses naturally handle actuator saturation, making the scheme suitable for low-cost embedded implementations.