This paper investigates interference-cancellation schemes at the receiver, in
which the original data of the interference is known a priori. Such a priori
knowledge is common in wireless relay networks. For example, a transmitting
relay could be relaying data that was previously transmitted by a node, in
which case the interference received by the node now is actually self
information. Besides the case of self information, the node could also have
overheard or received the interference data in a prior transmission by another
node. Directly removing the known interference requires accurate estimate of
the interference channel, which may be difficult in many situations. In this
paper, we propose a novel scheme, Blind Known Interference Cancellation (BKIC),
to cancel known interference without interference channel information. BKIC
consists of two steps. The first step combines adjacent symbols to cancel the
interference, exploiting the fact that the channel coefficients are almost the
same between successive symbols. After such interference cancellation, however,
the signal of interest is also distorted. The second step recovers the signal
of interest amidst the distortion. We propose two algorithms for the critical
second steps. The first algorithm (BKIC-S) is based on the principle of
smoothing. It is simple and has near optimal performance in the slow fading
scenario. The second algorithm (BKIC-RBP) is based on the principle of
real-valued belief propagation. It can achieve MAP-optimal performance with
fast convergence, and has near optimal performance even in the fast fading
scenario. Both BKIC schemes outperform the traditional self-interference
cancellation schemes with perfect initial channel information by a large
margin, while having lower complexities.