If I understand your question correctly, you are dealing with two unrelated problems. One is how to pass multiple pairs of fastq files (from the same sample but from different lanes) to bwa. The other problem is how to make snakemake accept multiple files in input linked to a single wildcard in output.

For the first issues, you can concatenate fastq files as @dr_bantz suggests but this could mean creating copies of possibly massive files. I've been dealing with this issue recently and I wrote a script to concatenate and interleave multiple pairs of fastq files which can then be piped to bwa (or anything that accepts interleaved files from stdin). Script is here catInterleaveFastq.

For the second problem I think something along these lines would work, probably it could be even simplified (see snakemake Functions as Input Files):

def get_fq1(wildcards):
    # code that returns a list of fastq files for read 1 based on *wildcards.sample* e.g.
    return sorted(glob.glob(wildcards.sample + '*_R1_001.fastq.gz'))

def get_fq2(wildcards):
    # code that returns a list of fastq files for read 2 based on *wildcards.sample*, e.g.
    return sorted(glob.glob(wildcards.sample + '*_R2_001.fastq.gz'))

rule bwa_map:
    input:
        fq1= get_fq1,
        fq2= get_fq2,
    output:
        bam= '{sample}.bam'
    shell:
        """
        catInterleaveFastq.sh -1 {input.fq1} \
                              -2 {input.fq2} \
        | bwa mem -p genome.fa - \
        | ... possibly pipe to "samtools sort"  etc \
        > {output}
        """

Hope this helps...

EDIT: Add -p to bwa command to enable detection of interleaved input.