Adaptive network-aware FeLAA LBT strategy for fair uplink FeLAA-WiFi coexistence

Figure A1 shows the discrete-time Markov chain proposed by this work. This model is based on the LAA specification [2]. The two chains in each Markov chain group represent the ICCA and ECCA procedures with the corresponding backoff stage mp, respectively. If K = 1, which means the device use CWtp,max for transmission for only one time, the Markov chain will consist of mp groups sorted according to the CW value in descending order. Otherwise, when K > 1, the end of this Markov chain will add K − 1 groups of which all the CW size are Wmp . Let pl be the collision probability of the FeLAA UE. If we adopt the short notation: P{s(t) = Ic(m), a(t) = i|s(t) = Ic(m), a(t) = i + 1} = P{Ic(m), i|Ic(m), i + 1} , the nonzero one-step transmission probabilities formulas of this Markov chain can be given by Eq. (A1)  P{Ic(m), i|Ic(m), i+ 1}=1− pl, i ∈ [0, N−1];m ∈ [0,mp +K − 1], P{Ic(m), N |Ie(m− 1), 0}=pl, m ∈ [0,mp +K−2], P{Ie(m), k|Ie(m), k + 1}=1, k ∈ [0,Wm−2];m ∈ [0,mp +K − 1], P{Ie(m), k|Ie(m− 1), 0}= pl(1−(1−pl) N ) Wm , k ∈ [0,Wm−1];m ∈ [1,mp +K − 1], P{Ie(0), k|Ie(mp +K − 1), 0}= pl(1−(1−pl) N ) Wl . k ∈ [0,Wl−1]. (A1)