Designing wireless sensor networks should meet appropriate parameters such as quality of service (QoS) defined by different users. The variable physical conditions of the environment, processing and transmission power limitations and limited communication capabilities are the most important obstacles that influence QoS parameters such as throughput, delay, reliability and network lifetime. The most important problems that directly have destructive effects on these parameters are hidden and exposed node problems. These problems extremely decrease throughput and increase delay and power consumption in the network. In this paper, we propose a new and efficient method that relies on clustering, grouping and subgrouping strategy to create smaller zones and avoid collisions. By allocating different transmission powers to different nodes depending on their hierarchy level in the network, the mentioned problems decrease and as a result power consumption decreases due to reduction in the retransmissions. This mechanism has been simulated by OPNET on IEEE802.15.4/Zigbee Manuscript profile