On the Decoder Error Probability of Rank Metric Codes and Constant-Dimension Codes

Rank metric codes and constant-dimension codes (CDCs) have been considered for error control in random network coding. Since decoder errors are more detrimental to system performance than decoder failures, in this paper we investigate the decoder error probability (DEP) of bounded distance decoders (BDDs) for rank metric codes and CDCs. For rank metric codes, we consider a channel motivated by network coding, where errors with the same row space are equiprobable. Over such channels, we establish upper bounds on the DEPs of BDDs, determine the exact DEP of BDDs for maximum rank distance (MRD) codes, and show that MRD codes have the greatest DEPs up to a scalar. To evaluate the DEPs of BDDs for CDCs, we first establish some fundamental geometric properties of the projective space. Using these geometric properties, we then consider BDDs in both subspace and injection metrics and derive analytical expressions of their DEPs for CDCs, over a symmetric operator channel, as functions of their distance distributions. Finally, we focus on CDCs obtained by lifting rank metric codes and establish two important results: First, we derive asymptotically tight upper bounds on the DEPs of BDDs in both metrics; Second, we show that the DEPs for KK codes are the greatest up to a scalar among all CDCs obtained by lifting rank metric codes.