Comparison of different Broadcast Schemes for Multi-Hop Wireless Sensor Networks

In this paper, we present the performance of different broadcast schemes for multihop sensor networks based on mathematical modeling. In near future many applications will demand multicast (Broadcast) communication feature from the sensor networks. This broadcast feature does not use virtual carrier sensing but relies on physical carrier sensing to reduce collision. For this paper, we analyze the different broadcast schemes for multihop wireless sensor networks and also calculated the achievable throughput.

💡 Research Summary

The paper addresses a critical gap in the design of multi‑hop wireless sensor networks (WSNs) where broadcast communication is increasingly required for applications such as environmental monitoring, emergency alerts, and distributed control. Traditional IEEE 802.11‑style MAC protocols rely heavily on virtual carrier sensing (VCS) – i.e., RTS/CTS exchanges – to mitigate hidden‑terminal collisions. However, broadcast frames cannot carry RTS/CTS, so VCS offers no protection and collision probability skyrockets in dense sensor deployments. To overcome this limitation, the authors focus on physical carrier sensing (PCS) as the primary collision‑avoidance mechanism and evaluate several broadcast schemes that differ in how they use PCS, back‑off, probabilistic suppression, and channel diversity.

System model – The authors consider N sensor nodes uniformly distributed over a two‑dimensional area. Each node has the same transmission power and a sensing radius R. Time is slotted, and nodes follow a CSMA/CA‑like protocol: they sense the channel, decrement a back‑off counter, and transmit when the counter reaches zero. Broadcast means a single transmission intended for all one‑hop neighbors. The traffic model assumes each node generates broadcast packets according to a Poisson process with rate τ.

Broadcast schemes evaluated

1. Pure PCS – Immediate transmission once the channel is sensed idle. This baseline has the lowest protocol overhead but suffers severe collisions as node density grows.
2. Adaptive Back‑off PCS – Nodes estimate current network load (e.g., by measuring idle‑slot frequency) and dynamically adjust the back‑off window size. Larger windows are used under high load to spread transmission attempts.
3. Probabilistic Suppression – Before transmitting, a node flips a biased coin with probability p (0 < p < 1). If the outcome is “suppress,” the node defers its broadcast to the next slot. This stochastic throttling reduces simultaneous transmissions without requiring explicit load estimation.
4. Multi‑Channel Broadcast – The available spectrum is divided into K orthogonal channels. Nodes randomly select a channel for each broadcast, thereby reducing the number of contenders per channel. The scheme incurs additional channel‑selection and synchronization overhead.
5. Hybrid PCS‑VCS – PCS remains the primary guard, but in high‑density regions a short RTS/CTS exchange is inserted to protect against hidden terminals. This hybrid approach aims to combine the low overhead of PCS with the hidden‑terminal mitigation of VCS.

Mathematical analysis – The authors construct a discrete‑time Markov chain for each scheme. The state space captures whether a node is idle, counting down, transmitting, or waiting for a back‑off reset. Transition probabilities are expressed in terms of three key parameters: node density λ (nodes per unit area), sensing radius R, and transmission attempt probability τ. The probability of a successful broadcast, (P_s), is derived as

\