Improved upper bounds on the packet error probability of slotted and unslotted DS/SS systems

We consider a direct-sequence spread-spectrum multiple access communication system with convolutional coding and hard decision decoding. We calculate the packet error probability and throughput of this system. Previous bounds on packet error probability have relied on the worst case assumption of phase and chip synchronous interference. We present new bounds on packet error probability and throughput for the case of both slotted and unslotted systems. These new bounding techniques are based on two advances: an improved Chernoff bound on the error event probability of a convolutional code, and the use of moment space techniques. Numerical results indicate that these new bounds on packet error probability improve on previously reported bounds by more than an order of magnitude. We also examine the problem of choosing the optimum code rate which maximizes throughput. We compare the optimum code rate which results from the bounding technique to the optimum code rate derived from an approximation technique. Although the bounding technique and approximation technique yield very different results for throughput, the resulting choice of optimum code rate is similar.