Path planning is an essential capability for autonomous mobile robot navigation. Taking inspiration from long-range navigation in animals, a neuromorphic system was designed to implement waypoint path planning on place cells that represent the navigation space as a cognitive graph of places by embedding the place-to-place connectivity in their synaptic interconnections. Hippocampal place cells, along with other spatially modulated neurons of the mammalian brain, like grid cells, head-direction cells and boundary cells are believed to support navigation. Path planning using spike latency of place cells was demonstrated using custom-designed, multi-neuron chips on examples and applied to a robotic arm control problem to show the extension of this system to other application domains. Based on the observation that varying the synaptic current integration in place cells affects the path selection by the planning system, two models of current integration were compared. By considering the overall path execution cost increase in response to an obstruction in the planned path execution, reduced spike latency response of a place cell to simultaneously converging spikes from multiple paths in the network was found to bias the path selection to paths offering more alternatives at various choice points. Application of the planning system to a navigation scenario was completed in software by using a place-cell based map-creation method to generate a map prior to planning and co-opting a grid-cell based path execution system that interacts with the path planning system to enable a simulated agent to do goal-directed navigation.