Software R&D management: commercial VS academic needs

Hello everyone,

I'm a project manager in a research institute in Russia. Our team comprises ~20 developers.
In the last 4 years, our main amusement was outsourcing software R&D projects for IT companies.

On the one hand, we are supposed to quickly (4-6 months for a project) deliver proof-of-concept prototypes with measured accuracy and performance, as well as to devise demo web-applications.
On the other hand, as an academy, we are interested in fundamental researching, i.e., trying different approaches to the same task, familiarizing with new technologies, improving overall R&D skills, submitting papers and patents, and so on.
Moreover, there is a constant need to maintain a steady personnel training process, just because about a half of our team are undergraduates and PhD students.

Due to lack of experience (all team members are <30) and absence of scientific advising, we constantly face a number of issues:
1. Poor project management & time scheduling: pursuing the "gold plate", we start to rush before every milestone
2. Minor additions to state-of-the-art approaches due to short project terms: poor scientific output, no contributions to the research field (papers, patents)
3. Lack of motivation: researchers suffer from short project terms, developers - from the need to quickly implement poor-quality prototypes
4. Poor inter-project communication: we have no time/resources to coordinate activities in adjacent projects, so the manager of several projects often becomes a single point of failure
5. Poor educational process: students involved in commercial projects have no time for fundamental learning and are forced to rush all the time

So the question is whether some resources (books/papers) exist that suggest possible solutions to our problems?
Today we feel ourselves stable and mature enough to try some commonly used methodologies for managing software prototypes R&D process and product lifecycle.

Any tips/clues/ideas are appreciated.

Tnx in advance!