Securing Data in Vehicles: Privacy-Preserving Frameworks for Dynamic CAV Environments

. 2026 Feb 19;26(4):1326. doi: 10.3390/s26041326

Algorithm 1 Federated Averaging. The K Clients are denoted as k; B, E, and η Represent the Local Minibatch Size, Number of Local Epochs, and Learning Rate, Respectively

1: initialize

w_{0}

2: for each round t = 1, 2, … do
3: m ← max (C · K, 1)
4: Select a random subset of m clients from the total set of
currently connected and eligible CAVs
5:

S_{t} \subseteq {1, 2, \dots, K}

6: for each client k ∈

S_{t}

in parallel do
7:

W_{t + 1}^{k}

← ClientUpdate (k,

W_{t}

)
8:

W_{t + 1}

←

\sum_{k = 1}^{K} \frac{n k}{n} W_{t + 1}^{k}

9: ClientUpdate (k, w): ‖Run on client k
10: β ← (split

P_{k}

into batches of size b)
11: β = {

b_{1}, b_{2}

,…}
12: for each local epoch i from 1 to E do
13: for batch b ∈ β do
14: w ← w − η▽ℓ (w; b)
15: return w to server