1. Everybody creates a new RSA keypair, signs it with the old private key and publishes it.
2. Everybody creates their new key pair
3. Everybody uses the new public key as message and creates the hash and signature padding for each of the Step 1 RSA keys. Then he blinds and signs them with his old keypair and publishes them.
4. Everybody textbook signs the blinded messages with his RSA private key from step 1. They can see which are valid, because they were signed with the original key.
5. Everybody unblinds his own public key messages

Now those new public keys are valid that have a valid signature of everybody's step 1 RSA key.

One quick idea. I didn't analyze it thoroughly, especially concerning active attacks by the participants.

1. Everybody creates a new RSA keypair, signs it with the old private key and publishes it.
2. Everybody creates their new key pair
3. Everybody uses the new public key as message and creates the hash and signature padding for each of the Step 1 RSA keys. Then he blinds them. And finally he signs them with his old keypair and publishes them.
4. Everybody textbook signs the blinded messages with his RSA private key from step 1. They can see which are valid, because they were signed with the original key.
5. Everybody unblinds his own public key messages

Now those new public keys are valid that have a valid signature of everybody's step 1 RSA key.