The purpose of this study is twofold. First, this study proposes a cost-effective LoRa gateway to improve bandwidth utilisation to achieve optimal network throughput for LoRa networks from a hardware design perspective. Secondly, this study creates a design for adaptive and autonomous algorithm to allocate bandwidth while meeting dynamic throughput demands from a network management perspective. This study conducted actual experiments to evaluate the network throughput capacity of a LoRa gateway based on the SemTechTM SX1301 transceiver chipset (SX1301 gateway). This study also addresses the limitations of the SX1301 gateway by proposing a gateway to improve bandwidth utilisation to increase network throughput. Knowing different packet sizes have an impact on network throughput, this study also set up a series of actual experiments related to the network throughput of the proposed gateway with 40 test combinations based on four different packet sizes and ten configurable bandwidth options. These test results were gathered and analysed to establish the throughput threshold for four different packet size ranges, and then to design an adaptive and autonomous algorithm for dynamic bandwidth allocation without human intervention. Based on the theoretical throughput capacity, the proposed gateway has an average improvement of throughput capacity of 57.73%, compared to the SX1301 gateway. The significance of the proposed adaptive algorithm is its capability for monitoring of the network usage constantly and then allocating the bandwidth on demand in an autonomous, agile and scalable fashion.