In this chapter, we consider the local MGOR scenario as in the example in Figure 8.1. Assume node S, i.e., the sender, is forwarding a packet to a remote destination D. S can transmit the packet at k different rates R1, R2, …, Rk. Each rate corresponds to a communication range, within which the nodes can receive the packet sent by S with some non-negligible probability, which is larger than a threshold, e.g., 0.1. The available next-hop node set 
(1 ≤ j ≤ k) of node S under a particular transmission rate Rj is defined as all the nodes in the communication range of S that are closer to D than S. We denote the nodes in 
as 
, where 
. In a similar way to geographic routing (Karp and Kung 2000; Lee et al. 2005; Seada et al. 2004), we assume S is aware of the location information of itself, its one-hop neighbors and the destination D. Define the packet advancement as 
in Equation (8.1), which is the Euclidian distance between the sender and destination (d(S, D)) minus the Euclidian distance between the neighbor 
and destination (
).
Figure 8.1 Node S is forwarding a packet to a remote destination D with transmission rate Rj. Reproduced by permission of © 2009 IEEE.
Then at each rate Rj, each node in 
is associated with one pair, 
, where 
is the data packet reception ratio (PRR) from node S to 
. Note that for different data rates, the PRR from node S to the same neighbor may be different. Let 
denote the forwarding candidate set of node S at rate Rj, which contains the nodes that participate in the local opportunistic forwarding. Note that here 
is a subset of 
, while in the existing pure opportunistic routing schemes (Biswas and Morris 2005; Zorzi and Rao 2003), 
.
The multirate GOR (MGOR) procedure is as follows: node S decides a transmission rate Rj, and selects 
based on its knowledge of 
(
's and 
's); then it broadcasts the data packet to the forwarding candidates in 
at rate Rj after detecting the channel is idle for a while. Candidates in 
follow a specific priority to relay the packet—that is, a forwarding candidate will only relay the packet if it has received the packet correctly and all the nodes with higher priorities failed to do so. The actual forwarder will become a new sender and suppress all the other potential forwarders in 
. When no forwarding candidate has successfully received the packet, the sender will retransmit the packet if retransmission is enabled. The sender will drop the packet when the retransmissions reach the limit. This procedure iterates until the packet arrives at the destination.
In this chapter we use a contention-based MAC protocol like 802.11 and apply a compressed slotted acknowledgement mechanism similar to that in Zubow et al. (2007) to coordinate the relay priority among the candidates, which is described as follows. After sensing the channel has been idle for a DIFS (distributed interframe space), the sender broadcasts the data packet at the selected rate. In the header of the packet, the intended MAC addresses of the forwarding candidates and the corresponding relay priorities are identified. If the first-priority candidate receives the packet correctly, it broadcasts an ACK with a delay of SIFS (short interframe space) after the successful data reception. The ACK is used for informing the sender of the data packet reception as well as suppressing lower priority candidates from forwarding duplicated copies. If the first-priority candidate does not receive the packet, it just remains silent. For the second-priority candidate, it sets a waiting period of 2TSIFS − Trx/tx after it received the data packet correctly, where TSIFS and Trx/tx is the time duration of SIFS and radio receive/transmit status turnaround delay, respectively. If within the waiting period, it detects a transmission emerged (e.g. a significant signal strength increase) in the channel, the ACK packet is considered as sent. Then it just drops the received packet. On the other hand, if no transmission emergence is detected, the second-priority candidate concludes that the highest prioritized candidate did miss the data packet. So the second-priority candidate will turn around its radio from receiving status to transmitting status and send out the ACK with 2TSIFS delay after it received the packet. Generally, the ith-priority (i > 1) candidate that receives the data packet will set a waiting period as i × TSIFS − Trx/tx after the data packet reception. If it detects a transmission emerging in this period, it will suppress itself from forwarding the packet; otherwise, it will send out an ACK at i × TSIFS to claim its reception. In Section 8.2.4, we will further elaborate on the impact of reliability of this ACK technique on the performance of OR.