IMPROVED UPPER-BOUNDS ON THE PACKET ERROR-PROBABILITY OF SLOTTED AND UNSLOTTED DS SS SYSTEMS/

We consider a direct-sequence spread-spectrum multiple access communic ation 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 c ase 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 p robability of a convolutional code, and the use of moment space techni ques. Numerical results indicate that these new bounds on packet error probability improve on previously reported bounds by more than an ord er of magnitude. We also examine the problem of choosing the optimum c ode rate which maximizes throughput, We compare the optimum code rate which results from the bounding technique to the optimum code rate der ived from an approximation technique. Although the bounding technique and approximation technique yield very different results for throughpu t, the resulting choice of optimum code rate is similar.